Final Report Summary - EURHEALTHAGEING (European ResearcH on DevElopmentAL, BirtH and Genetic Determinants of Ageing)
Executive Summary
Concept and Objectives: EurHEALTHAgeing was a multidisciplinary project linking studies of early developmental processes with those on longevity and ageing The study was highly integrated and included genetic, epigenetic, biochemical, bioinformatic and statistical approaches that tackles the complex problem of unravelling the molecular basis of aging and is translating it into biomarker development. The project is subdivided into 7 scientific and technological workpackages a management WP a dissemination WP and an ethics WP
Main results
WP1 - A GWAS on birthweight (bw) was carried out in TwinsUK and replicated in HCS.. Common genetic variants identified by the EGG consortium to be associated with birthweight were selected for WP2. An epigenomewide scan on birthweight was carried out in discovery and replication samples from identical twins discordant for birth weight. Metabolomic screens using the Metabolon and Biocrates panels on birthweight were performed in twins and the top 25 metabolites forwarded to for de novo replication in the other cohorts (WP5). NextGen Seq Association Analysis. Borderline hit of interest: NBPF1 Implicated in a number of developmental and neurogenetic disorders
WP2 carried out replication genotyping of the top hits from twins and singletons and this was analysed as part of WP6. In addition it developed XworX, a workflow based, cloud enabled analysis platform which can be downloaded from http://www.xworx.org/#!download/ch3t. In order to guarantee scalability, flexibility & reusability of pipelines all tools are submitted with default inputs for easy and user friendly handling.
WP3 carried out the assays on 96 methylation probes for epigenetic replication and 396 450k Infinium (genomewide) methylation arrays in twins and singletons all of which were analysed in WP6.
WP4 optimized and technically validated candidate biomarkers measuring specific post-translational modifications(PTMs). After preliminary analyses the following biomarkers were chosen: Following three biomarkers were measured: C1M (Type I collagen degradation, measures a MMP-generated fragment of type I collagen) ; C3M (Type III collagen degradation, measures a MMP-generated fragment of type III collagen) ; and VICM (Citrullinated vimentin degradation, measures a MMP generated citrullinated vimentin fragment) . C3M was modestly associated in twins and singletons with birthweight. VICM was associated with birthweight in singletons only. Analysis with age related traits showed an association of VICM with cardiovascular disease risk, of C1M and C3M with osteoarthritis and gout.
WP5 developed and optimized the assays to measure 25 selected metabolites and carried out the measurements
WP6 carried out the statistical analysis and pathway analysed the data generated from WP2-3-4-5. The significant pathways from genetic and epigenetic analysis are involved in blood pressure, obesity and height which are in agreement with published findings linking early growth factors with adulthood disease. Joint analyses of singletons and twins found consistency in associations with birthweight within singletons for all three technologies (genetics, epigenetics, metabolomics) but only for metabolomics between twins and singletons. Genetic results in singletons are consistent an a genetic risk score for birthweight correlates with measures of metabolic disease and blood pressure in adults Epigenetic analyses resulted in only one gene being nominally (but not Bonferroni) significant between twins and singletons. This is the MIA3 gene which is known to be associated with cardiovascular disease. Metabolomics results in all three cohorts are consistent, and a metabolomic signature in adults is seen related to birthweight (AUC= 0.6). The metabolites associated with birthweight are also associated with blood pressure and insulin resistance in adults. .
In WP7, an integrated platform was implemented in order to help the researchers to compare the data derived from EurHealth studies with those that come from other studies based on the correlations between early life events and ageing outcomes linking datapoints to methylation, metabolomic, genetic and post-translational modification (PTM) results. All tasks have been completed as planned in the original workplan.
WP8 A number of dissemination activities, including a summer school / symposium in the last year of the project, were carried out as part of the project. Seventeen reviewed publications some of which received considerable media attention were produced. Proper running of the project was ensured by the management WP9. The ethical issues relating to the project was followed by WP10 and an external ethics advisor supervised that all procedures were in accordance to current EU legislation and guidance.
Project Context and Objectives:
Project context and objectives
There is growing evidence that developmental and biological disruptions during the early years of life can lead to age-related diseases. By 2050 the number of people in the EU aged 80+ age group will increase by 170% with resulting health care spending increases. Healthy life expectancy must evolve parallel to these changes if cost increases are to be constrained. (The impact of ageing on public expenditure - DG ECFIN 2006, p. 133). Effective health policy across the lifecycle can support healthy aging of population. Traditional risk factors from adult life or even genetic predictors derived from studies of longevity are able to only predict a small fraction of variation in healthy ageing between individuals. Therefore there is a pressing need to identify of early life events and risk factors that determine health outcomes in later life.
EurHEALTHAgeing was a multidisciplinary project linking studies of early developmental processes with those on longevity and ageing by combining use of cutting edge technologies- post-translational modifications, DNA methylation, metabolomics and genomics with unique cohorts with multiple ageing phenotypes. We used four key European study cohorts with rich early life data and age-related health outcomes. We will explore a unique twin dataset, two birth cohorts with extensive maternal, pre- and peri-natal data and a population based cohort of subjects in their 60s-70s. By analysing specific age related metabolomics, epigenetic (methylation and post-translational modification) data and correlating it to early life events, genetics and ageing outcomes we will provide major insights into developmental processes that influence longevity and ageing. By focussing on molecular mechanisms involved in both early life events and ageing we found metabolic and biochemical markers as well as lifecourse pathways that are relevant in both early development and adult life.
These data can be now tested as biomarkers of ageing that reflect the role of early development on ageing; potentially identifying pathways for therapeutic intervention when the process is still reversible.
Figure 1 shows a schematic representation of how pre- and peri-natal molecular mechanisms will be used for understanding and assessing predictive ability for changes occurring later in life affecting health status during the entire lifetime of individuals.
Figure 1. Link between early developmental processes and healthy ageing in EurHEALTHAgeing, (CHD=coronary heart disease, T2D=type 2 diabetes)
see attached document for figures.
The underlying hypothesis of the project was that developmental plasticity modulated by processes such as methylation and post-translational modifications is influenced by early life events and determines risk of age related diseases later in life.
We analyses peri-natal factors compared to four key areas of age-related diseases for which information is available in the EurHEALTHAgeing study cohorts.
The Specific Objectives were :
WP1 Data Mining and Sample Logistics. The main objectives to this WP are: (1) to identify genetic variants involved in birthweight and other early post natal determinants of healthy ageing and test the variants identified by ours and other studies against the ageing traits . This includes mining in house and published GWAS data and NextGen Sequencing to identify rare variants that influence birthweight and ageing phenotypes (2) Carry out an epigenome wide scan (EWAS) on birthweight and a metabolomic screen on birthweight in the TwinsUK samples to identify potential methylation probes and metabolites involved in early development and aging to be replicated in the additional cohorts (3)To organise the shipment of DNA and biospecimen samples from the clinical collections to the SMEs for genetic, methylation, metabolomic and post-translational modification testing
WP2 Technology coordination and Genetics. The objective of this WP are: (1) To define epigenetic candidate marker panels based on the results from KCL EWAS . To technically confirm candidate markers using Targeted Deep Bisulfite Sequencing (TDBS). To design and set up qPCR assays for confirmed methylation markers and transfer experimentally validated assays to WP3 for high throughput screening. (2) To define candidate SNPs markers based on the results of primary GWAS studies, design and set up nano-liter qPCR assays for high throughput SNP markers and perform high throughput SNP screening and transfer data to WP6.
WP3 Epigenetic Assays and Validation: To set up standard operating procedure including quality control for methylation analysis in high throughput on microfluidic qPCR platform, and analyze samples on mehtylation markers identified by WP1 and assays developed by WP2. To perform Illumina 450k analyses on a subset of the cohort samples.
WP4 Post Translational Modifications & Validation: The objective of this WP is assess the role of post-translational modifications known to influence rates of tissue ageing in MZ twins discordant for weight at birth and to replicate the PTMs thus identified in additional cohort samples
WP5 Metabolomic Assays and Validation: This WP will carry out development and validation of the mass spectrometric assays and perform high throughput quantification of 25 pre-selected candidate metabolites (WP1, KCL) in 5000 serum samples by mass spectrometry
WP6 Statistical Modelling. The main goal of this WP is identify trajectories leading to healthy aging by incorporating early life events, methylation, post-translational modification, metabolomic and genotypic data
WP7 Data Integration and Health Benefit Evaluation. The main objective of this WP is to generate an extensive database of the correlations between early life events and ageing outcomes linking datapoints to methylation, metabolomic, genetic and post-translational modification (PTM) results, allowing future comparison with data derived from other studies. Results will be useful in tests for healthy ageing.
WP 8: Management & WP 9: Dissemination.These work packages concentrate on the management and dissemination of the EurHEALTHAgeing Project
WP 10. Ethics The objectives of this WP are to ensure compliance with EU and member state ethical legislation and to oversee the ethical concerns of the EurHEALTHAgeing project
Project Results:
WP1 Data Mining and Sample Logistics.
Task 1.1. A GWAS on birthweight (bw) was carried out in TwinsUK and replicated in HCS. The results from this work were published in Twin Research in 2014. Common genetic variants identified by the EGG consortium to be associated with birthweight have been selected for WP2. Table 1 includes the list of markers identified to be validated by WP2. In addition to SNPs associated with birthweight other SNPs selected from GWAS on height, blood pressure and obesity were also chosen to be tested.
Table 1: list of markers selected for further de novo replication with birthweight in the EurHEALTH cohorts
Task 1.2. An epigenomewide scan on birthweight was carried out in discovery and replication samples from identical twins discordant for birth weight. DNA from white blood cells was obtained in 24 (discovery) and 21 (replication) BW-discordant female monozygotic (MZ) twin-pairs, and 86 unrelated female subjects from the TwinsUK cohort. Methylation profiles were characterized using the Infinium Human Methylation 450 BeadChip Kit. Following quality-control checks, within-twin-pair methylation differences were correlated to BW differences. Meta-analysis across the discovery and replication twin datasets was performed to identify birth-weight differentially methylated regions (BW-DMRs). Altogether, 12 CpGs (within/near genes HSF1, THADA, BRSK2, MAFF, FNINP1, RAPGEF6, MPP1, ENDOD1, TMEM126A, SLC9A3, and RAB3A) were differentially methylated in the twin-pair meta-analysis (P < 1×10-8). Of these, 54.5% share the same direction of association between methylation and BW in the unrelated dataset of 86 female subjects. We further validated the direction of BW-methylation associations at several of these regions using methylated DNA immunoprecipitation followed by deep sequencing (MeDIP-seq) in the 21 BW-discordant MZ twin-pairs.. The top 350 probes associated have been selected for assay development by WP2 and validation by WP3. Details on the probes that have passed QC for further development are presented under WP2. A manuscript is being drafted with these results. (see WP6)
Task 1.3. Metabolomic screens using the Metabolon and Biocrates panels on birthweight were performed in twins and the top 25 metabolites forwarded to WP5. The results of the Metabolon analyses have been published and the open access article is available at: http://ije.oxfordjournals.org/content/early/2013/06/29/ije.dyt094.long
Task 1.4 Shipments of DNA and biospecimen samples from the clinical collections to the SMEs for genetic, methylation, metabolomic and post-translational modification testing have been carried out.
Task 1.5 Work on the NGS pipeline was carried out. A report with QC aligment and call report was submitted to the EU portal.
Task 1.6 NextGen Seq Association Analysis. Single point has one significant hit but no information on biological relevance. Borderline hit of interest: GLI2: zinc finger (7.41E-06) Mutations in this gene may contribute to sterol accumulation and atherosclerosis, and have been observed in patients with sitosterolemia.SKAT (collapsing of rare variants)- a number of hits but most significant orfs, FAM proteins (no biological information). Borderline hit of interest: NBPF1 Implicated in a number of developmental and neurogenetic disorders
WP2. Technology coordination and Genetics
Task 2.1
T2.1 comprises the validation of candidate marker panels selected from data of previous GWAS and MEDIP-Seq. The selection of the marker panels was headed by KCL and was done in accordance with the other partners of the consortium. Electronic lists of epigenetic as well as SNP candidate markers were defined and forwarded to AIT for assay design. The lists were extended by markers identified by thorough literature research. Thus the lists are based on robust statistical analyses and already identified possible marker candidates. AIT compiled a SOP for the preliminary preparation of the samples to ensure that proper amounts of DNA were shipped to the corresponding project partners.
Results: Based on the previous genome wide GWAS and MEDIP-Seq data statistical robust candidate marker panels have been selected which indicates a significant difference in birth weight (low vs. high), birth height (small vs. large), blood pressure and BMI. List with candidate markers for both epigenetic and SNP validation studies have been forwarded to AIT. The assay design is accomplished and a SOP for the shipment of the samples to the corresponding partners of the consortium has been drafted. Shipment of the samples to TATAA and AIT for the epigenetic and SNP analysis is already completed.
Task 2.2
The objective of T2.2 is the development of workflow based tools for facilitated assay design and data analysis. All tools (TBS Primer Design; TBS data analysis; assay design for MSRE and MSP qPCR; SNP data analysis tool; Epigenetic qPCR data analysis) developed within that project are made available on XworX, a workflow based, cloud enabled analysis platform which can be downloaded from http://www.xworx.org/#!download/ch3t. In order to guarantee scalability, flexibility & reusability of pipelines all tools are submitted with default inputs for easy and user friendly handling.
Each tool consists of a short description at the beginning, as well as a more detailed description with additional information at the end. Tools can be started by clicking the green triangle at the XguI window and the results will be shown at the XtablE or XvieW window, respectively. An overview of all bioinformatics tools can be found at http://www.xworx.org/#!tools/ceap.
Figure 2.1: Overview of the developed bioinformatics tools in EurHEALTHAging project.
Figure 2.2: Access to the tool specific webpages for each bioinformatics tool.
In addition, by using the dropdown options under the header entry “Tools” (Figure 3) tool specific webpages for each bioinformatics tool providing information about: Input, Output, Tool version, Programming language, License, Screencast link, Download link for the basic tool code, Runtime and Reference information, Recommendations for tool usage and contact information can be found.
Summary: We have finished the
• TBS primer design tool
• MSRE-and MSP based qPCR design tool
• SNP qPCR design tool
• TBS data analysis pipeline
• SNP qPCR analyses pipeline
• Epigenetic qPCR analysis pipeline
All workflows are now available as tools on the XworX platform, which can be downloaded from "http://www.xworx.org/#!download/ch3t". Furthermore, we added a tool specific webpage for each bioinformatics tool providing additional information and screencasts.
T2.3 Technical validation of candidate markers using Targeted Bisulfite Sequencing (TBS)
– preliminary report in 1st period
– accomplished in 2nd reporting period
Summary of progress towards objectives
Aims
1. Assay design for candidate markers of relevant group comparisons
2. Technical confirmation of candidate markers using Targeted Bisulfite Sequencing (TBS).
3. Selection of technically validated epigenetic candidate sequences for the design of high-throughput MSRE based (or MSP-) qPCR.
To support the 450k methylation data with additional epigenome wide information, a selected panel of twins pairs (n=24) were analysed by Illumina’s HumanMethylation450 BeadArrays (450k BeadChip). This was a major improvement to the initially stated project proposal, as the use of genome wide methylation BeadChips massively improves the available data and allows robust statistics. The 450k Bead Chip interrogates 485577 single CpG’s distributed over the whole genome and returns the methylation state of the individual cytosine in terms of percentage. DNA of the 48 samples underwent quality assessment and was quantified using measurements based on DNA intercalating fluorescence dyes (PicoGreen). All samples passed the quality check and were deaminated according to the recommendations of Illumina. The deamination procedure with sodium bisulphite converts unmethylated cytosines into uracil by removing the amino group from the nucleotide, whereby methylated cytosines remain unaffected. To ensure sufficient amounts of DNA whole genome amplification was performed with subsequent enzymatically fragmentation of the DNA, followed by several washing procedures to ensure highly purified DNA. In the next step the processed DNA was hybridized onto the 450k BeadChips. Finally the BeadChips were scanned on an Illumina iScan. Data preprocessing and analysis was done according to Illumina’s recommendations and BRB Array Tools, this included transformation of signal intensities into beta-values, quantile normalization, inference statistics, hierarchical clustering and ROC analysis. The genome wide data was analysed with respect to the different birth weight and body height of the twins. The 450k data was also analysed for any correlations with the Holloway – Isle of Wight study.
Results
Statistical evaluation of the data revealed the highly discriminative methylation pattern between twins with low and high body weight/height. However the absolute differences of approximately 10% in methylation intensities were smaller than expected.. Based on the available data derived from methylation arrays a selection of markers was nominated in accordance with the participating partners of WP2 for further confirmation on a large independent sample cohort. A technical validation of the MEDIP-Seq and BeadChip results by TBS was planned. But the small methylation differences (~10%) derived from the 450k BeadChips are hardly detectable by qPCR based methods. Therefore, we started a technical validation run with the qPCR based method, to evaluate which methylation differences are capable by those methods. The technical validation by TBS was executed after that confirmation experiments done by the qPCR based methods. Consequently, we conducted the confirmation of the 450k BeadChip data by two independent methods. This was an improvement compared to the initial stated project plan.
450 k vs MSRE qPCR
The methylome of 24 twin pairs has been analysed by Illumina’s 450k BeadChip technology, which is an improvement compared to the initial project proposal. The use of state-of-the-art technology enabled us to perform a cost efficient whole genome analysis of the methylome. This yielded high statistical power and a robust selection of features for subsequent independent testing using independent samples and an independent technology. A technical validation of the 450k BeadChip by targeted bisulfite sequencing (TBS) was conducted on the same 24 twin pairs. Assays for TBS were designed using the XworX platform. The basic for the assay design were the 90 MSRE-qPCR assays included in the final marker panel (see T2.4). 63 out of the 66 TBS assays have been successfully tested and optimized in the lab. Methylation values received by TBS and by 450k BeadChips showed a high concordance (R=0.77-0.98) and confirmed the marker selection, which derived from the 450k BeadChip.
T2.4 Design, setup and validation of methods for nano-liter MSRE qPCR for epigenetic high through-put testing
- accomplished in 2nd reporting period
Aims:
1. Selection of technically validated epigenetic candidate sequences for the design of high-throughput MSRE based (or MSP-) qPCR.
2. Assay design for candidate markers of relevant group comparisons
For the methylation assays 305 epigenetic candidate markers with differentially methylated loci (DML) were nominated based on the data of the previous analysis (T2.1 and additionally conducted 450k experiments). The MSRE-primer-design tool (“Xprimer_MSRE”) – developed in T2.2 – was used to design and retrieve primer sequences for all 305 candidate markers. qPCR assays were successfully designed for all 305 epigenetic candidate markers. Independently of the number of sequences for which assays should be designed, the “Xprimer_MSRE” tool takes approximately 37s for the primer design of each input sequence. Consequently, “Xprimer_MSRE” supports the user to design the best possible primer pair for the given DNA sequence. Subsequently it was possible to evaluate the suitability of the designed assays in silico by evaluating the presence of SNPs in the primer sequence, the presence of repeats and the number of given cut sites for the methylation sensitive restriction enzymes (HpaII, Hin6I, AciI, HpyCH4IV).
Subsequently, each MSRE qPCR assay which passed the in silico evaluation was subjected to MIQE conform primer testing in single as well as in multiplexed approaches using DNA isolated from peripheral blood. All primers were tested on three different days with three different DNA samples (1 male, 1 female and DNA from 1 commercial available cell line [MCF7]) in serial dilution series (start concentration 10 ng/reaction, dilution factor: 4; calibration points: 4).
Minimum selection criteria for MSRE qPCR assays were as follows:
• Discriminative power (measured on the Illumina 450k BeadChip): gene regions with highly significant p-values (p<0.01) and at least a 10% difference in methylation signal intensities between the groups.
• Regions including SNPs which could interfere with restriction digestions and/or primer sequences were avoided
• At least two cut sites for the MSREs HpaII, Hin6I, AciI, HpyCH4IV.
• Long stretches of repeats and assays completely on repeats were avoided, due to the danger of reduced qPCR performance.
• Similar properties across the different assays (including primer length [18-28 bp], amplicon length [65-130 bp] and melting temperature [63-67°C]) to enable multiplexed high throughput qPCR assays.
The successfully tested qPCR assays were applied to the same twin DNA, which was also used for the 450k BeadChip experiments to confirm the findings from the BeadChips by qPCR based methods. That is a deviation to the initially stated project proposal, which describes that the confirmation will be done by TBS only. The reason for that were the small differences in methylation state between the envisioned groups.
For technical confirmation by the MSRE based qPCR the results from the 450k BeadChip have been reproduced in the same sample set. A member of TATAA visited AIT by the end of September 2013 for training purposes. The training at AIT covered the whole workflow. The large scale sample analysis was performed after that training at the labs of TATAA in Göteborg.
T2.5 GWAS/SNP replication
In previous studies GWAS experiments on a large panel of twins were conducted. Those GWAS experiments served as basis for the SNP marker selection within the EurHEALTHAgeing project. The marker selection was done in accordance with KCL, and additional SNPs with potential impact on birth weight and birth height were selected from literature. The final panel of selected SNPs contained 94 SNPs with impact on birth height and birth weight. SNPtype assays were provided by Fluidigm (San Francisco). Fluidigms high throughput SNPtype genotyping assays were based on a specific target amplification to warrant the enrichment of the targets of interest, even in the case that only small amounts of DNA are available. Enrichment was realized by a multiplexed preamplification with primers specific for the targets of interest (STA and REV primers). The preamplification was done in a 5 µl scale in a 384 well format. 1 µl of the 5 µl preamplification was diluted 1:100 and was sufficient for several hundred analyses. The readout was based on the amplification with allele specific primers (Fwd primer), amplifying in dependence of the presence of the corresponding allele either the wildtype (homozygote), the SNP (homozygote) or both (heterozygote). The detection consists of a simple endpoint measurement. The readout was done on micro-fluidic chips covering 24 assays x 192 samples, making the platform the ideal choice for high throughput testing.
Summary: 94 SNPs were selected and subjected to SNPtype assay design (Figure 13). Seventy-eight SNPtype assays were successfully designed of which 72 were passed the quality criteria in several test runs and showed high concordance with SNP data derived from microarrays. Those 72 assays were applied to 3168 samples, which were provided by the University of Southampton. Sample testing has been finished in calendar week 37. Data processing is still in progress and will be made available to KCL and UOULU by the end of September at the latest.
Three different control DNAs isolated from peripheral blood were included in each Fluidigm Chip. Concordant call rates for the control DNAs were >98%. D2.5 due in M30 was delivered 11 month earlier than stated in the project proposal. After finalizing the measurements the received data was forwarded to the bioinformatics section of Southampton as stated in the project proposal.
Different significant assocations between the investigated SNPs and the parameters body weight, BMI, sysBP and diaBP have been identified.
WP3- Epigenetic Assays and Validation:
The main objective for work package 3 was to analyse samples provided from work package 1 for the methylation markers identified in work package 2 using the optimised procedures developed and to generate genomewide methylation data for both singletons and twins.
Task 3.1: Reagents for pre-amplification and high-throughput qPCR have been optimised and compared to current leading commercial reagents. Especially the pre-amplification mastermix require special conditions to be compatible with the reagent carryover from the restriction enzyme step prior to pre-amplification.
The results from the validation were reported in deliverable D3.2. Most assays worked well and all of them, except ABL1, will be used for the analysis of the samples from cohort studies. ABL1 which did not work at all was replaced by a negative control, which was lacking in the original design. Due to the difference in Cq values between samples for the same assay, for the continuation of this project both cleaved and uncleaved samples will be analysed together to be able to correctly determine the degree of methylation.
In addition to bias the most important property of the pre-amplification step is the reproducibility.. Both number of assays with a standard deviation above 0.5 and the average standard deviation for all assays are similar or better for the developed mastermix than for the leading commercial mastermixes.
The Fluidigm Biomark instruments used for qPCR analysis of pre-amplified material is not compatible with regular SYBR Green I based qPCR mastermixes. The most commonly used dye for the Biomark is EvaGreen. A mastermix based on EvaGreen was developed and compared to one of the leading EvaGreen based commercial mastermixes. Figure 3.2 shows the correlation between the developed mastermix and a commercial mastermix commonly used for the Biomark. In general the Cq-values for a qPCR reaction starts to loose reproducibility above Cq 30-34, but for the Biomark Cq-values are normally at least 5 cycles lower which means that reproducibility starts to increase at cycle 25-29. This can be seen in Figure 3.3 where the reproducibility of the qPCR replicates shows a good correlation until about Cq 25, while at higher Cq-values the qPCR replicates start to differ.
The correlation between the developed and commercial mastermixes shows a good correlation, especially at lower Cq-values. At higher Cq-values the uncertainty of the Cq determination increases the variability of the data. Further optimisation was performed to increase signal level without losing reproducibility. The optimised reagents were sent to AIT for evaluation with the methylation assays developed in WP2.
Task 3.2: The standard operating protocol (SOP) describing all steps for the DNA methylation analysis of samples from the project using quantitative polymerase chain reaction (qPCR) was finalised as planned. The SOP is described deliverable D3.1. The steps of the analysis consists of restriction enzyme cleavage of the DNA using methyl sensitive restriction enzymes, pre-amplification of the digested DNA samples and qPCR analysis of the pre-amplified samples. The main instrument for this method is the high-throughput platform BioMark from Fluidigm. The assays and procedure used in the SOP are validated in Task 3.1 and deliverable D3.1.
Task 3.3: During the second period 8832 sample reactions (from 4416 cleaved and uncleaved samples, respectively) obtained from WP1 were analysed by methylation sensitive restriction enzyme (MSRE) quantitative real-time PCR (qPCR) according to the standard operating procedure (SOP) described in deliverable D3.1. In order to achieve the best result for each sample regardless of any possible concentration differences of the samples each sample was divided into two parts where one part was digested using the MSREs and the other was treated with the same buffers and temperatures without the MSREs. The level of methylation was then based on the difference in Cq between the two parts and calculated assuming 90% efficiency, based on the assay validation in deliverable D3.2.
Among the 96 assays were 6 controls. Two of them, JUB and IRF4, are positive controls, i.e. they have no site for the restriction enzymes and should have the same Cq values for both digested or undigested samples. One, TBP, targets an un-methylated region with a site of the enzymes which should be completely cleaved in the digested samples. Another control assay, SNRPN, targets an imprinted region which means the ideal methylation level is 50% and consequently the difference in Cq between the digested and undigested samples should be 1 (or 1.1 for 90% efficiency of the assay). One assay, XIST, targets the X-chromosome and therefore differentiates between men and women, where women’s values should look like SNRPN and the men’s should look like JUB or IRF4. As mentioned in deliverable D3.2 one of the control assays (ABL1) did work at all and that was therefore not analysed in the samples, but instead replaced with water as a negative control, blank.
The results of the analyses are reported in deliverable D3.3 and deliverable D3.3 Appendix 1. The raw data was transferred to partner 6, AIT, for calculation of methylation degrees. These were further transferred to the cohort partners for statistical analysis.
Task 3.4 Genomewide DNA methylation quantification using the Infinium 450k
DNA was extracted from whole peripheral blood (stored in EDTA tubes) by standardized salting out methods. Genome-wide DNA methylation levels were measured using Illumina Human Methylation 450K array (Illumina). In short, in two batches, samples (500ng of DNA per sample) were first bisulfite treated using the Zymo EZ-96 DNA-methylation kit (Zymo Research, Irvine, CA, USA). Next, they were hybridized to the arrays according to the manufacturer’s protocol. The methylation percentage of a CpG site was reported as a beta-value ranging between 0 (no methylation) and 1 (full methylation).
Quality control of the samples was done with Genome Studio and all samples had a call rate >99%, and complete bisulphite conversion.
Quality control of the probes was done based on the detection p-value calculated with Genome Studio. Probes with a detection p-value of more than 0.01 in more than 1% of the samples were excluded.
Methylation beta values were generated using the 450K DataProcessing pipeline, which is based on the pipeline of Tost and Touleimat (2012, Epigenomics), with background correction. We excluded individuals with leukemia or received chemotherapy. Lastly, we excluded samples whose correlation with our reference population was r<0.80
Quality Control for Ilumina 450k data
Prior to analysis, we cleaned the data for potential outliers by first clustering the subjects from each dataset by hierarchical clustering based on the pairwise Pearson’s correlation to produce the correlation heatmap and boxplot for the methylation distribution in each subject. Secondly, we correlated the first two principal components (PC) to the covariates to identify systematic batch effects. The plate, position on plate, and bisulfite conversion levels are associated with methylation levels, so these three batch effects and age are included as covariates in all the analysis. In all datasets, methylation probes that mapped to multiple locations within 2 mismatches, with missing values, mapped to Y chromosome, and with detection p-value>0.05 are removed. Finally, a total of 475,529 probes including probes on the X chromosomes are carried through to the birthweight-DMR analysis. 396 samples were analysed using the 450k and were forwarded to participant #1 for epigenome wide analysis of birthweight and other traits Results are summarized under WP6.
WP4. Post Translational Modifications & Validation
Deliverable D4.1. The main objective of the first delivery was to optimize and technically validate candidate biomarkers measuring specific PTMs.The main tasks of the delivery D4.1 were do optimized and validate the citrullinated vimentin assay (VICM) and the nitrosylated type III collagen assay (CO3-nys), as well as doing some preliminary test of the assays in relevant samples.
It was decided that VICM was an interesting assay to measure in samples this showed some predictive value for structural progression in ankylosing spondylitis. However the CO3-nys assay showed little biological relevance and it was decided that this part of the project should be discontinued. Instead other PTM assays were proposed as being more interesting for the project and therefore focus will change to those markers in the coming period.
Deliverable D4.2. The main tasks of the deliverable D4.2 were to test CO3-nys, C3M and VICM in a sample cohort collected from an elderly population. The purpose of the sample testing was to pre-validate the assays before measurement in the cohorts allocated to the project. Furthermore an additional task was added to the WP. As CO3-nys failed initial evaluation, new biomarker candidate should be proposed. First we tested the two markers in a cohort of 65 elderly and investigated whether the markers were associated with age, sex or BMI. There was no significant correlation between the markers and those parameters (table 4.2).
Next we measured whether levels of the markers would be higher in individuals with above normal degree of inflammation (hsCRP>5) as compared to individuals with normal degree of inflammation (hsCRP<5).
Deliverable D4.3. The main tasks of the delivery D4.3 were to develop and test a range of novel collagen turnover marker of type I, IV, V and VI collagen. Furthermore, we wanted to continue the evaluation of the VICM assay.
The levels of MMP generated fragments of type I (C1M) and III (C3M) collagen were significantly elevated in serum from lung cancer patients as compared to controls (Figure 4.1A-B) with an average increase of 8-fold and 2-fold, respectively. These findings indicate that altered collagen turnover is ongoing in lung cancer. The levels of citrullinated and MMP generated fragments of vimentin (VICM) were significantly elevated (13-fold on average) as compared to controls (Figure 4.1C). These data together with previously presented data indicate that these markers may be markers of an unhealthy phenotype and may contribute in the prediction of disease events.
Figure 4.1.
We have prepared and tested the several proposed markers. We found that C1M, VICM and C3M were robust and showed good technical performance, and felt comfortable to measure those in the cohort samples of the project.
Deliverable D4.4. The main tasks of the delivery D4.4 were to test a selective panel of biomarkers in several studies; 1) The UK-Twin study and 2) The HCS study. Following biomarkers were measured:
- C1M (Type I collagen degradation, measures a MMP-generated fragment of type I collagen)
- C3M (Type III collagen degradation, measures a MMP-generated fragment of type III collagen)
- VICM (Citrullinated vimentin degradation, measures a MMP generated citrullinated vimentin fragment)
All biomarker kits were produced under GMP and validated before run in the cohort samples following internal SOPs at Nordic Bioscience. This includes setting limits for reruns and failed sample detection.
The biomarkers tested showed only a weak correlation with birthweight, both C1M and VICM are appear associated with birthweight in the twin cohort. Also VICM was associated with CVD.
Deliverable D4.5. The main tasks of the delivery D4.5 was to test two of the best biomarkers from D4.1 to D4.4 C1M and C3M, in the UK-Twin study (N=2500). Tested for relationship with birth weight and various age related co-morbidities (GOAL cohort, mean age 69 years)
Gout and OA are the two most prevalent rheumatological conditions, both of them show a strong association with age and both are usually accompanied by cardiometabolic comorbidities. In the case of gout a 30-70% increased risk in CVD mortality has been reported. We investigated the role of the two collagen degradation markers in these two conditions. Association between C3M with OA after adjustment for age, and BMI gave an OR =66.5 [6.9-643] p <0.0003 and with gout OR =82.8 [9.9-695] p <2.4 x 10[-5].
However C1M was strongly associated with prevalence of COPD and C3M appears to be associated with prevalence of epilpetic seizures. The increased risk seen was also consistent with the negative association seen with birthweight according to DOHaD. There was no association to other co-morbidities such as Heart disease, hypertension, osteoporosis and type II diabetes (table 4.3).
Table 4.3
Trait C1M
C3M
OR 95%-CI P-value OR 95%-CI P-value
COPD 3.46 1.49-8.03 0.0004 1.48 0.44-4.98 Ns
Epilepsy 1.32 0.79-2.22 Ns 2.01 1.01-4.01 0.048
In summary: We developed and tested several biomarkers measuring PTMs. We found a strong relationship between the biomarkers C1M and C3M, an unhealthy outcome. This may indicate that these biomarkers may act as diagnostic and prognostic tools.A technician, as well as part of a scientist, were recruited as part of this project. These will continue their employment of Nordic Bioscience to further explore the diagnostic potential of the investigated biomarkers.
WP5 Metabolomic Assays and Validation
Task 5.1. Definitiono f 25 metabolites. These were derived from the analysis of nominally significant metabolites associated with birthweight as described in WP1. The list of these metabolites was forwarded to SL to develop the necessary assays.
Task 5.2: Development and validation of mass spectrometric MRM-assays for 25 metabolites (Duration: months 6-24)
For each metabolite, collision-induced dissociation (CID) was optimized and the most characteristic fragment ions were selected for building up the MRM-assays. Ultimate goal was the inclusion of as many metabolites as possible in one chromatographic run. Some metabolites could be sensitively detected without derivatization (method 1). However, for increased sensitivity and chromatographic separation, most metabolites had to be derivatized (butylated; method 2). Serum was used as sample matrix. Assay-validation and sample measurement comprised following steps:
Selection of materials (standards, instruments)
Name Company Ordering number
Standards Method 1
Urea PlusOne 17-1319-01
Uric acid Sigma U2625
DL-2-Hydroxybutyric acid sodium salt Alfa Aesar A18636
C-Glycosyl tryptophan RIKEN (personal agreement with RIKEN and Shino MANABE, scientist) -
DL-p-Hydroxyphenyllactic acid Aldrich H3253-100MG
Hexadecanedioate Aldrich 177504-1G
Standards Method 2
Amino acid standard acidic and neutral Sigma A6407
DL-Kynurenine Sigma 61250
N-Acetylglycine Aldrich A16300
H-Glu (Leu-OH)-OH Bachem G-1950
H-Glu (Val-OH)-OH Bachem G-2015
Phenylacetyl-L-Glutamine Santa Cruz sc-212551
N-Acetylthreonine Chem-Impex #03262
2-Methylbutyroyl carnitine Medical Isotopes 5242
Acetyl-L-carnitine Dr. Herman J. ten Brink -
Dodecanoyl-L-carnitine Dr. Herman J. ten Brink -
Isovaleryl-L-carnitine Dr. Herman J. ten Brink -
Oleoyl-L-carnitine Aldrich 597562
Propionyl-L-carnitine Dr. Herman J. ten Brink -
Glutarylcarnitine TRC G597600
Succinylcarnitine TRC S688830
Pyridoxic acid Sigma P9630
Deuterated internal standards
DL-Aspartic acid-d3 Cambridge Isotope Laboratories DLM-832
Glycine-d2 Cambridge Isotope Laboratories DLM-1674
L-Glutamic acid-d3 Cambridge Isotope Laboratories DLM-3725
L-Kynurenin-d4 Buchem -
L-Tyrosine (Ring-d4) Cambridge Isotope Laboratories DLM-451
N-Acetyl-d3-threonine-2,3-d2 Medical Isotopes D5843
Phenylacetyl-d5 L-glutamine Medical Isotopes D30083
Hexadecanoyl-L-carnitine-d3 Dr. Herman J. ten Brink -
Acetyl-L-carnitine-d3 Dr. Herman J. ten Brink -
Isovaleryl-L-carnitine-d9 Dr. Herman J. ten Brink -
Propionyl-L-carnitine-d3 Dr. Herman J. ten Brink -
2-Methylbutyroyl N-(methyl-d3)-carnitine Medical Isotopes D32092
NSK-B-G mix standard Cambridge Isotope Laboratories NSK-B-G
Pyridoxic acid-d2 buchem 4-PYRA-D2-001
Urea 15N2 Cambridge Isotope Laboratories NLM-233
Uric acid 15N2 Cambridge Isotope Laboratories NLM-1697
--Hydroxybutyric acid-d6 sodium salt Lipomed GHB-752-NA-100
Testosteroneglucuronide-d3 NARL 97-000056
Hexadecanedioate-d28 CDN Isotopes D-5186
Other
Methanol HPLC-grade Merck 1.06035.1000
Water HPLC-grade Biosolve 23214102
Acetonitrile HPLC-grade Biosolve 01204101
3N HCl/1-butanol Fluka 87472-50ml-F
V-vials Phenomenex AR0-3740-13 1000/PK
Bördelk./Crimp Caps N11 Macherey-Nagel 70231
LC-MS/MS unit Thermo Scientific, TSQ Vanatge TQU02485
LC-MS/MS unit Thermo Scientific, TSQ Quantum -
Column Dionex 063189
Column Zorbax 966967-906
Sample preparation
100 µl of human serum were mixed with 10 µl of internal standard. For protein precipitation, 850 µl of methanol were added, the mixture was vortexed two times for 5 seconds and centrifuged for 3 minutes at 10°C and 14.000 rcf. 200 µl of the supernatant were transferred into a new V-vial and dried under vacuum for 1 hour at 45°C. The dried material was re-dissolved in 100 µl of 20% methanol and 6 metabolites were measured directly using LC-MS/MS (method 1). 700 µl of the supernatant were transferred into a second V-vial and dried under vacuum for 2 hours at 45°C. For derivatisation, 100 µl of 3N HCl/1-butanol were added, the mixture was vortexed, incubated at 60°C for 7.5 minutes and dried under vacuum for 45 min. at 45°C. The dried material was re-dissolved in 100 µl of 20% acetonitrile/0.1% formic acid and 19 metabolites were measured using LC-MS/MS (method 2).
Calibration: To generate calibration curves for method 1, six different concentrations (standards 1, 2, 3, 4, 5 and 6) of the requested substances were mixed with 10 µl of internal standard in new V-vials. The mixture was dried under vacuum for 1 hour at 45°C. The dried material was re-dissolved in 100 µl of 20% methanol and the 6 metabolites were directly measured using liquid chromatography mass spectrometry (LC-MS/MS). As an example, the calibration curve of C-glycosyl tryptophan is shown in figure 1. To generate calibration curves for method 2 eight different concentrations (standards 1, 2, 3, 4, 5, 6, 7 and 8) of the requested substances were mixed with 10 µl of internal standard in new V-vials. The mixture was dried under vacuum for 2 hours at 45°C. For derivatisation, 100 µl of 3N HCl/1-butanol were added, the mixture was vortexed, incubated at 60°C for 7.5 minutes and dried under vacuum for 45 minutes at 45°C. The dried material was re-dissolved in 100 µl of 20% acetonitrile/0.1% formic acid and 19 metabolites were measured using LC-MS/MS. As example, the calibration curve of propionylcarnitine is shown in figure 2.
Figure 1: Calibration curve of C-glycosyl tryptophan. On the x-axis C-glycosyl tryptophan concentration [ng/ml] and on the y-axis the area ratio (ratio between signal of standard and internal standard) are shown.
Figure 2: Calibration curve of propionylcarnitine. On the x-axis propionylcarnitine concentration [ng/ml] and on the y-axis the area ratio (ratio between signal of standard and internal standard) are shown.
LC-MS/MS
Method 1For direct analyses, an LC-MS/MS instrument of the TSQ Vantage series (Thermo Scientific), a reversed-phase C-18 column (Dionex Acclaim PolarAdvantage II, 4.6 x 50mm, 3µm) and following chromatographic conditions were used: 100% water/0.2% formic acid (1 min, isocratic), linear gradient to 100% methanol/0.1% formic acid (3 min), 100% methanol/0.1% formic acid (2.5 min, isocratic) and linear gradient back to 100% water/0.2% formic acid (1.5 min).
Method 2For analyses of derivatised samples, an LC-MS/MS instrument of the TSQ Quantum Access Max series (Thermo Scientific), a reversed phase C-8 column (Zorbax XDB, 4.6 x 75mm, 3.5µm) and following chromatographic conditions were used: 15% acetonitrile and 85% water (0.1 min), linear gradient to 100% acetonitrile (6.5 min), 100% acetonitrile (1.5 min, isocratic) and linear gradient back to 15% acetonitrile and 85% water (1 min).
Validation: Each standard was measured three times and the limits of quantitation and detection (LOQs and LODs, see table below) were determined using ValiData3.00. To evaluate matrix effects, these standards were mixed with 100 µl of human control serum from a healthy male volunteer and also analysed. No matrix interferences were observed. To guarantee repeatability, human control serum was analysed for 5 times on three different days. Repeatability was confirmed. To test the precision of the method, human control serum was measured ten times in a row. Robustness was tested by varying three steps during sample preparation, namely vortexing of samples (one time/30 seconds vs. two times/5 seconds), centrifugation (6 minutes instead of 3 minutes), and derivatisation time (10 minutes instead of 7.5 minutes). Each modification was analysed three times. Results showed that the methods were precise and robust.
Table 1: LOQs and LODs for the 25 metabolites
Metabolite LOQ LOD
Method 1
Urea 10 µg/ml 3 µg/ml
Uric acid 7 µg/ml 2 µg/ml
2-Hydroxybutyric acid 879 ng/ml 240 ng/ml
3,4-Hydroxyphenyllactic acid 9 ng/ml 2 ng/ml
C-glycosyltryptophane 32 ng/ml 9 ng/ml
Hexadecanedioate 10 ng/ml 3 ng/ml
Method 2
Glycine 1 µg/ml 260 ng/ml
Tyrosine 1579 ng/ml 438 ng/ml
N-Acetylthreonine 13 ng/ml 4 ng/ml
Kynurenine 9 ng/ml 3 ng/ml
N-Acetylglycine 270 ng/ml 75 ng/ml
Aspartic acid 1690 ng/ml 470 ng/ml
Glutamic acid 2 µg/ml 445 ng/ml
Phenylacetylglutamine 87 ng/ml 24 ng/ml
--gluamylleucine 81 ng/ml 23 ng/ml
--glutamylvaline 97 ng/ml 27 ng/ml
Pyridoxic acid 3 ng/ml 1 ng/ml
Acetylcarnitine 64 ng/ml 18 ng/ml
Propionylcarnitine 7 ng/ml 2 ng/ml
2-Methylbutyroylcarnitine 5 ng/ml 1 ng/ml
Isovalerylcarnitine 2 ng/ml 1 ng/ml
Succinylcarnitine 3 g/ml 1 ng/ml
Glutarylcarnitine 4 ng/ml 1 ng/ml
Dodecanoylcarnitine 5 ng/ml 1 ng/ml
Oleoylcarnitine 9 ng/ml 2 ng/ml
Task 5.3: High throughput quantification of 25 pre-selected candidate metabolites in 5000 serum samples by mass spectrometry (Duration: months 26-32)
After method development and validation, all ca. 5000 samples were analysed and results delivered to project partners It was agreed that all samples are returned after analysis and this was accomplished by month 35.
The analysis of these results in the three cohorts combined is presented in WP2 (task 6.2)
WP6 Pathway analysis and statistical modelling
Task 6.1 Pathway Analysis
Our pathway analysis approach is based on collective effects of the groups of genes interlinked by functional relationships (see [1]). To maximise the advantage of pathway analysis, we concentrated on the "group behaviour" of genes, their ability to interact and pre-existing annotation placing the genes into the same biological pathway, linking to the same cellular function. The benefits of the use of pathway and ontological analyses of genomics data have been extensively presented in the past [2-3].
Single genes that do not map into any statistically significant pathway will still be considered significant if reproducible and independently validated in more than one of the approaches used (methylation, NGS, GWAS, metabolomics) as they may happen to be in large pathways where the rest of the genes have no effect, and so the pathway is not significant, yet being highly significant in themselves. However, for our analysis pipeline we will leave such genes out since our approach is based on collective effects of the groups of genes interlinked by functional relationships, which is inapplicable to some genes lacking information on function, regulation and interaction with other genes.
Analysis of biological pathways will be performed using The Database for Annotation, Visualization and Integrated Discovery (DAVID ) v6.7 (4)
Genes through GWAS
List of genes associated with birthweight
ANGPT4 CCNL1 HMGA2 CDKAL1
CPOE CENPM CALCR 5q11.2
CDK2 ADCY5 LDLRAD2 LCORL
GRB10 CPEB3 RALA NKAIN2
OSBPL5 PARG HLA KY
REG1B GALNT13 HHIP
RUNX2
GO TERM Pathway P_Value
OMIM_DISEASE Many sequence variants affecting diversity of adult human height
2.80E-03
Genes Through Epigenetics
A disease enrichment analysis was performed using web-based gene set analysis toolkit by comparing our top 51 hits to the entrez gene database. Several disease-associated genes were found and the table below lists the top associated diseases (at least 3 genes are associated with the disease, adjust P-value< 0.01) and genes. Additionally, we found two genes, GJA4 and ITGA4, previously reported to be associated with infertility (adj. P-value= 0.0085) and NEUROG3 and DIO2 to associate with Type 2 diabetes (adj. P-value= 0.0363).
List of diseases associated top birthweight-differentially methylated regions in Twins.
Disease Gene Adj. P-value
Stroke;
Stroke NOS;
Cerebral Infarction PRKCH, GJA4, LTBP2, SORCS2 0.0011
0.0011
0.0011
Subarachnoid Hemorrhage PTPRN2, RGS12, SORCS2 0.0014
Diabetes Mellitus;
Endocrine disturbance NOS;
Endocrine system Diseases;
Endocrine disorder NOS PTPRN2, BACH2, NEUROG3, DIO2 0.0022
0.0032
0.0032
0.0032
Autoimmune Disease PTPRN2, ZMIZ1, BACH2, ITGA4 0.0032
Type I Diabetes Mellitus PTPRN2, BACH2, NEUROG3 0.0032
Genetic predisposition to disease LSP1, GJA4, PTBP2, ZMIZ1, BACH2 0.0039
Infarction PRKCH, GJA4, SORCS2 0.005
Metabolic diseases PTPRN2, BACH2, NEUROG3, DIO2 0.0063
Skin Diseases (genetic) LSP1, ZMIZ1, MTA1 0.0076
*Adj. P-value: P-value adjusted by multiple tests
Figure showing a summary of the biological process categories of the 39 genes involved. Among these genes, 14 of them are involved in metabolic processes, and 13 are involved in developmental processes.. Top 39 genes involved biological process
Combined Genetic and epigenetic molecular pathways
P_Value
GOTERM_BP_FAT activation of adenylate cyclase activity
6.10E-03
GOTERM_BP_FAT positive regulation of adenylate cyclase activity
6.40E-03
GOTERM_BP_FAT positive regulation of cyclase activity
6.60E-03
GOTERM_BP_FAT positive regulation of lyase activity
7.00E-03
OMIM_DISEASE Many sequence variants affecting diversity of adult human height
9.60E-03
GOTERM_BP_FAT regulation of adenylate cyclase activity
1.80E-02
GOTERM_BP_FAT regulation of cyclase activity
1.90E-02
GOTERM_BP_FAT regulation of cAMP biosynthetic process
2.00E-02
GOTERM_BP_FAT regulation of lyase activity
2.00E-02
GOTERM_BP_FAT regulation of cAMP metabolic process
2.00E-02
GOTERM_BP_FAT regulation of cyclic nucleotide biosynthetic process
2.30E-02
GOTERM_BP_FAT regulation of nucleotide biosynthetic process
2.30E-02
GOTERM_BP_FAT regulation of cyclic nucleotide metabolic process
2.40E-02
GOTERM_BP_FAT regulation of nucleotide metabolic process
2.50E-02
Discussion of Significant Pathways
• activation and positive regulation of adenylate cyclase activity
Any process that increases the frequency, rate or extent of adenylate cyclase (AC) activity that is an integral part of a G-protein coupled receptor signaling pathway. This term can be used to annotate ligands, receptors and G-proteins that lead to activation of adenylate cyclase activity within a signaling pathway.
• positive regulation of lyase activity
Any process that modulates the frequency, rate or extent of lyase activity, the catalysis of the cleavage of C-C, C-O, C-N and other bonds by other means than by hydrolysis or oxidation, or conversely adding a group to a double bond. They differ from other enzymes in that two substrates are involved in one reaction direction, but only one in the other direction. When acting on the single substrate, a molecule is eliminated and this generates either a new double bond or a new ring.
• Many sequence variants affecting diversity of adult human height
• regulation of cAMP biosynthetic process
Any process that modulates the frequency, rate or extent of the chemical reactions and pathways involving the nucleotide cAMP (cyclic AMP, adenosine 3',5'-cyclophosphate).
• regulation of cyclic nucleotide biosynthetic and metabolic process
Any process that modulates the frequency, rate or extent of the chemical reactions and pathways resulting in the formation of cyclic nucleotides.
Discussion
The significant pathways are involved in blood pressure, obesity and height which are in agreement with published findings linking early growth factors with adulthood disease.
Adult height has been found to be inversely associated with mortality. Data from the Danish Medical Birth Register have shown that there is a strong positive association between birth weight and adult height; for subjects with low birth weight and that genetic and/or environmental factors operating both during the pre- and postnatal period may be responsible for the association between birth length and adult height [5]
There is a very clear link between adult blood pressure and birthweight with low birthweight being a significant risk factor for adult hypertension [6,7] and obesity [8].
The results presented here shed light into the molecular pathways in common between these two phenomena.
Furthermore, we find that there is concordance in the pathways if not in the individual genes identified by genetic and epigenetic approaches.
References
1. Ptitsyn AA, Weil MM, Thamm DH. Systems biology approach to identification of biomarkers for metastatic progression in cancer. BMC Bioinformatics. 2008 Aug 12;9 Suppl 9:S8.
2. Draghici S, Khatri P, Martins RP, Ostermeier GC, Krawetz SA: Global functional profiling of gene expression. Genomics 2003, 81(2):98-104
3. Manoli T, Gretz N, Grone HJ, Kenzelmann M, Eils R, Brors B: Group testing for pathway analysis improves comparability of different microarray datasets. Bioinformatics 2006, 22(20):2500-2506
4. Systematic and integrative analysis of large gene lists using DAVID Bioinformatics Resources. (2009) Nat Protoc. 4(1):44 -57
5. Sørensen HT, Sabroe S, Rothman KJ, Gillman M, Steffensen FH, Fischer P, Sørensen TI. Birth weight and length as predictors for adult height. Am J Epidemiol. 1999 Apr 15;149(8):726-9.
6. Sato R, Maekawa M, Genma R, Shirai K, Ohki S, Morita H, Suda T, Watanabe H. Final Height and Cardiometabolic Outcomes in Young Adults with Very Low Birth Weight (<1500 g). PLoS One. 2014 Nov 14;9(11):e112286. doi: 10.1371/journal.pone.0112286
7. McNamara BJ, Gubhaju L, Chamberlain C, Stanley F, Eades SJ. Early life influences on cardio-metabolic disease risk in aboriginal populations--what is the evidence? A systematic review of longitudinal and case-control studies. Int J Epidemiol. 2012 Dec;41(6):1661-82. doi: 10.1093/ije/dys190. Epub 2012 Dec 3.
8. Botton J, Scherdel P, Regnault N, Heude B, Charles, MA Postnatal weight and height growth modeling and prediction of body mass index as a function of time for the study of growth determinants. Nutr Metab. 2014;65(2-3):156-66. doi: 10.1159/000362203. Epub 2014 Nov 18.
Task 6.2. statistical modelling
Genetic and life-course analyses leading to healthy ageing – approach from early life
The intrauterine period is a vulnerable period of development. Any adverse environment can permanently change the body’s organ structure and function, expressed as an increased disease risk later in life. Studies show that variability in growth patterns in early life is associated with obesity, and other cardiovascular diseases in adulthood, but the genetic and environmental determinants of these processes are largely unknown. In EurHealthAgeing we aimed to identify genetic, metabolic and environmental factors associated with early growth in infancy and childhood and later metabolic outcomes in adulthood. The analyses using early childhood growth data were based on Northern Finland Birth Cohort 1966 (NFBC1966) and NFBC1986 data while the rest of the analyses e.g. on birth measures and ageing related outcomes used also TwinsUK and Hertfordshire Cohort Study (HCS). Genome-wide analyses (GWAS) on childhood growth included also other birth cohorts studies externally to EurHealthAgeing form Early Growth Genetics (EGG) consortium (as planned due to statistical power issues).
Our results multivariable modelling show that several maternal and paternal factors, such as socioeconomic status, height, smoking, parity and pre-eclampsia, have direct and independent associations with postnatal height growth patterns, some of which had their association mediated by size-at-birth variables. It was observed that an obesogenic environment in utero and during a child’s growth exerts a ‘programming’ effect on the glucose-insulin axis as well as cardiovascular risk factors in adolescence. We identified genetic variants (e.g. in FTO-gene) that showed an age-dependent association with adiposity in early childhood, while others clearly had their effect on adult adiposity mediated by early growth phenotypes. These analyses emphasise the clinical importance of early growth markers as it may inform public health policy aimed to improve the pre-pregnancy environment and also monitor infant growth during the first years of life (Das, thesis 2014, Das et al 2015). Our discoveries on FTO-gene led us to investigate further its’ function. These results demonstrate that this gene has a role in adipose tissue which modifies the response of white adipose tissue to high-fat feeding. Fto-deficiency in mouse model increases the expression of genes related to adipogenesis preventing adipocytes to become hypertrophic after high-fat diet (Ronkainen et al 2015, in press). Our results emphasize the importance of further work on molecular mechanisms.
For further statistical analyses, a birth weight lowering genetic risk score was calculated for Northern Finland Birth Cohort 1966 (NFBC1966) and Hertfordshire Cohort Study (HCS). The risk score was based on 17 of 22 SNPs formerly identified as being associated with birth weight (Horikoshi et al 2012, where NFBCs and HCS were included), reduction in the number of SNPs being due to availability in HCS. The association of the risk score with birth weight itself was validated in both cohorts. As birth weight itself is associated with various adult phenotypes, it was hypothesized that these same genetic factors affecting birth weight could be associated with health-related outcomes in later life; i.e. may explain partially the observed associations between foetal growth and adult phenotypes.
The first set of outcomes was adult adiposity measures: BMI, weight, hip circumference and waist circumference and in NFBC1966 also blood pressure. The results for adiposity measures showed significant negative association with the birth weight lowering risk score and all adiposity measures in the NFBC66 (n=4912). However, there was no evidence for any association in the HCS (n=2397). This could be due to the age difference in the two cohorts; the NFBC1966 measurements were done at the age of 31 years, while in the HCS the phenotypes in question were measured at much higher age (~ age of 60 years). The above analyses are not conducted before and show novel, slightly unexpected but plausible results. The detailed analyses are presented in the Appendix.
In NFBC1966, the effect of the birth weight lowering genetic risk score on adult blood pressure was also examined. Evidence for negative association between the risk score and systolic blood pressure was found, even after various adjustments. We did not find any statistically significant association for diastolic blood pressure. All above work is being prepared for publication .As the next step, the same risk score was calculated for Northern Finnish Birth Cohort 1986 (NFBC1986) and the association with the same set of phenotypes will be examined. As the data for NFBC1986 is measured at the age of 16, this will provide more information on the association at yet another time point compared to NFBC1966 and HCS.
Both NFBC1966 and NFBC1986 have data measured on multiple time points, ranging from maternal variables to the most recent 46-year study for NFBC1966 collected between 2012 and 2014. For these longitudinal datasets, structural equation models (SEM) have been constructed until the age of 31 years, available for the present study, in order to gain more depth for understanding the life-course association with the genetics and early-life variables and the later life phenotypes related to “healthy ageing”. This work will continue by using the latest data which should be available in 2015 for the analyses. Later on, a subset of these wider life-course models could be used in cohorts with no longitudinal data available. The latest completed analyses during the reporting period using SEM on 5198 participants from the NFBC1966 with data on birth weight, height and weight measurements until adolescence, systolic and diastolic BP at 31 years and several other covariates including genetic risk score for blood pressure (Figure 1). Construction of these final models includes multiple sets of analyses, very careful insight into the individual associations between the variables, and testing of the model performance (the baseline developmental work is described in the thesis by Marika Kaakinen in 2013). Negative direct effects of birth weight on adult systolic BP were observed (standardised regression coefficients: −0.08 (−0.14 to −0.03) in males and −0.04 (−0.09 to 0.01) in females, equalling −1.99 (−3.32 to −0.65) and −1.01 (−2.33 to 0.32) mm Hg/kg, respectively). Immediate postnatal growth was associated with adult BP only indirectly via growth later in life. In contrast, growth from adiposity rebound onwards had large direct, indirect and total effects on adult BP. Current body mass index was the strongest growth-related predictor of adult BP (0.36 (0.30 to 0.41) in males and 0.31 (0.24 0.37) in females, equalling 1.29 (1.09 to 1.48) and 0.81 (0.63 to 0.99) mm Hg/(kg/m2), respectively). As a summary, our path analytical approach provides evidence for the importance of both foetal growth and postnatal growth, especially from adiposity rebound onwards, in determining adult BP, together with genetic predisposition and behavioural factors. These analyses also give much more in-depth understanding about the interplay of different contributing factors and show how important it is to follow-up groups at risk from very early beginning of life. This kind of work is the starting point for effective primary prevention (Kaakinen et al 2014).
In addition to the above longitudinal analyses we have been working on environmentally and lifestyle related factors in adult life which may be causally related to ageing related conditions. One of these potential contributing factors is vitamin D status. Our results suggest that higher 25(OH)D leads to reduced risk of hypertension, providing support for important non-skeletal effects of vitamin D (Vimaleswaran et al. 2014).
Metabolomic data analyses (see Tables below)
Based on Menni et al paper (2013), 25 metabolites associated with ageing were identified and the data for these metabolic variables in TwinsUK (n NA), HCS (n NA) and NFBC86 (n=2394) was provided by Seibersdorf Laboratories. The association with these metabolites and birth measures (birth weight, birth length, ponderal index, head circumference) were examined. The analyses were stratified by sex. For birth weight, the female results were meta-analysed and 12 metabolites were identified being associated with birth weight.
After the meta-analyses, in order to examine the risk for a low birth weight, a birth weight metabolomic risk score for females was constructed in each cohort separately. This was done by treating the lowest tertile of birth weight vs. the rest as a binary dependent variable (value 1 if individual belong in the lowest birth weight tertile). Stepwise regression was used in order to find out if any of the 12 metabolites could be excluded from the risk score, nevertheless no support for excluding any variable was found by these stepwise methods. However, due to severe collinearity, one metabolite was removed from the final model, leaving 11 metabolites in the final risk score model. The AUC in a ROC curve was then calculated for each three cohorts. The next step in the metabolites analysis is to examine the relationship with the metabolites and the genetic risk score in NFBC1986. These works are being written up for publication. We have also participated in important methodological work related to assessing of multivariate gene-metabolome associations with rare variants. This work proposes a new statistical approach based on Bayesian reduced rank regression to assess the impact of multiple SNPs on a high-dimensional phenotype. Because of the method's ability to combine information over multiple SNPs and phenotypes, it is particularly suitable for detecting associations involving rare variants. The work demonstrates the potential of our method and compare it with alternatives using the Northern Finland Birth Cohort with 4702 individuals (Marttinen et al 2014).
Methylation data analyses
For DNA methylation, the data for 90 assays and 6 control assays was provided by TATAA for TwinsUK (n=1034), NFBC1966 (n=1521), NFBC1986 (n=1527) and HCS (n=269). The associations with the markers and birth weight were examined and meta-analysed for TwinsUK, NFBC1966 and NFBC1986 data, using a brief QC of scaling and quantile-normalising the assay values. Evidence for association was found for a CpG site near MIA3 gene where variants have been associated with cancer development and coronary artery disease.
Methylation data using Illumina 450K array (HM450K) data has recently become available for both NFBC1966 (n=816) and NFBC1986 (n=552) (in late 2014) as well as in HCS and TwinsUK. A thorough QC pipeline process is set up based on our experiences using this array. The first results show that almost all previously published top hits associated with maternal smoking during pregnancy (using cord blood DNA and HM450K; Richmond et al 2014) show also association in NFBC1966 in samples taken at age 31 years. These results are extremely valuable from public health point of view. The QC process for our HM450K data is further ongoing in order to provide usable data for epigenome-wide association analyses. The original plan was not to use HM450K arrays at larger scale because their usability was not yet established when the current study was designed. However, during the project’s time more evidence was gathered and it because evident that it will be feasible to carry on with larger scale analyses. Our ongoing work also approves HM450K’s value also in population based studies. In the future, the longitudinal life-course models described earlier could be appended with both metabolomic and epigenetic data. The analyses done until this point give excellent starting points in creating these models, and the wide variety of data enables creation of complex models which, given such a complex set of phenotypes as “healthy ageing”, are severely needed in order to understand the very nature of healthy ageing.
Overall comparison of results between datasets
Overall the analysis of genetic, epigenetic and metabolomics data showed consistent effects for singletons and twins for metabolites. A metabolomics profile in adult age reaches 60% of the area of under the curve for birthweignt. On the other hand, genetic factors are clearly different between singletons and twins, but are reproducible within singletons and are associated with blood pressure. Epigenetic analysis identified only two probes consistently associated with birthweight in both singletons and twins. The genetic risk score analyses with adiposity related adult outomes showed different patters by age of the cohort.
EPIGENETIC ANALYSES
EWAS for birthweight in Twins and singletons using 450k data
The genomewide methylation generated in Task 3.4 was used to perform epigenomewide scans in both twins and singletons
EWAS for birthweight in twins: Manhattan plot for birthweight EWAS in TwinsUK. Manhattan plot of bw-EWAS. Each green dot represents is a transformed –log10(p-value) value of one CpG site. Blue line indicates p-value of 10-4 and red line indicates Bonfferoni correction criteria P = 1.05×10-7. None of the hits reached genomewide significance.
List of top probes in the twins
IlmnID CHR Gene NAME β_Observatory β_Replication P-value
cg26174880 2 - -0.708 -0.646 8.27E-07
cg06699564 8 - -0.193 -0.852 4.95E-06
cg12165758 14 PRKCH -0.574 -0.677 5.71E-06
cg22145181 3 - -0.184 -0.842 8.73E-06
cg01324261 4 SCRG1 0.638 0.610 9.30E-06
cg14410072 9 C8G;FBXW5 -0.786 -0.450 9.90E-06
cg23366832 6 BACH2 0.459 0.702 1.94E-05
cg01510588 14 C14orf183 0.756 0.443 2.38E-05
cg12415687 7 PTPRN2 -0.588 -0.608 2.42E-05
cg21258821 13 KBTBD6 0.722 0.481 2.52E-05
cg26621897 X TMSB15A 0.363 0.739 2.74E-05
cg12961733 22 - -0.515 -0.653 2.88E-05
cg05630111 1 LASS2 -0.555 -0.620 3.24E-05
cg06866628 9 GTF3C5 -0.350 -0.738 3.38E-05
cg16280098 14 PABPN1 -0.667 -0.519 3.73E-05
cg15222563 19 TRAPPC5 0.534 0.625 4.00E-05
cg06062821 5 - 0.561 0.606 4.01E-05
cg02120071 2 - 0.486 0.655 4.20E-05
cg06973667 10 NEUROG3 0.473 0.663 4.22E-05
cg15323253 6 - 0.756 0.410 4.35E-05
cg17045635 10 ZMIZ1 0.640 0.536 4.39E-05
cg07171024 14 DPF3 -0.595 -0.575 4.40E-05
cg12846139 17 BAHCC1 -0.747 -0.415 4.83E-05
EWAS for birthweight in singletons: The same techniques were used to perform an EWAS on birthweight in the singleton cohorts. Again, no genomewide significant hits were found and very little overlap in the top 10000 probes with those from twins.
LIST OF TOP HITS IN SINGLETONS:
Gene CPG_ISLAND TargetID chr Cjr [psotopm_36 bw_pvalue bw_beta bw_se
cg15287987 7 117791566 2.47E-06 4.0791 0.7919
cg25967146 8 23773851 8.28E-06 -3.1752 0.6574
ODZ3 cg16793274 4 183951366 1.15E-05 3.1993 0.6746
DHX36 cg27297043 3 155517786 1.49E-05 -3.2903 0.7044
SEMA3A cg16346212 7 83662191 1.55E-05 2.9506 0.6332
CYP2C19 cg04189838 10 96513337 3.05E-05 -3.4565 0.7726
TAPBP N_Shore cg26315802 6 33389950 3.24E-05 3.2391 0.7268
IGBP1 Island cg19629755 69269900 3.34E-05 1.8673 0.4197
LOC100128811 Island cg06967120 10 25504265 3.54E-05 3.3740 0.7612
MLPH N_Shore cg17187163 2 238059746 4.06E-05 -3.3565 0.7639
BMP7 Island cg20340302 20 55274556 4.24E-05 -2.8304 0.6459
cg12804647 1 48243188 4.28E-05 -3.3038 0.7544
GPR25 N_Shore cg24550456 1 199108567 4.79E-05 -2.9578 0.6803
KRT83 Island cg16160105 12 51001317 4.79E-05 3.4751 0.7994
The lack of agreement in the genes found between twins and singletons suggest that epigenetic markers of birthweight, as are also genetic influences, are different in twins and singletons and should be analysed separately. This was confirmed by the meta anlaysis of the hits analysed by MSRE based (or MSP-) qPCR (see below)
A note of caution is that the EWAS results derive in total from less than 200 singleton samples and 296 twins. One of the starting assumptions of the project was that epigenetic effect sizes on birthweight would be much larger than genetic effect sizes and therefore even a few hundred individuals would be sufficient to identify consistent epigenetic effects. This has not proven to be the case and one important conclusion from this study is that larger discovery sample sizes are needed.
Epigenetic replication of initial discovery hits in data generated by MSRE based qPCR
Using the MSRE qPCER methylation data we combined data from all cohorts (total n= 3400) for the probes that passed QC that were generated by D3.3. The meta-analysis results for the association of candidate markers with birth weight (TwinsUK, NFBC1966, HSC and NFBC1986) are shown in the table below.
Marker P-value Beta SE
ABHD1 0.737085744 0.006759832 0.020135613
AGAP1 0.996476861 8.79047E-05 0.019907835
ASXL1 0.471239524 0.014506611 0.02013511
ATP9B 0.468891731 0.014522085 0.020050285
BAI3 0.509812045 0.013157687 0.019962198
BCAN 0.962405699 -0.0009521 0.020199531
BLCAP 0.888311968 -0.002792969 0.019887213
C11orf92 0.481901528 0.014070445 0.020007821
C1orf53 0.437383539 0.015521126 0.019985513
C21orf67 0.559641627 0.011721698 0.020092946
CD32 0.241915817 0.023543095 0.02011866
CRB2 0.664595767 0.008713903 0.02009771
DDB2 0.262007354 0.022535881 0.020091554
DIP2C 0.291600224 0.021045675 0.019955737
EIF2AK1 0.520164932 -0.012917211 0.020086133
EP4 0.273352979 -0.022011635 0.020095089
ESYT1 0.356925003 0.018509664 0.020092115
F5 0.43575647 0.015544673 0.019944952
FAM172A 0.112387259 0.031385014 0.019769416
FBXL21 0.514669563 0.01309028 0.020089941
FOXG1 0.433742262 -0.015721538 0.020083627
FOXQ1 0.314666838 0.020166048 0.020056183
HES7 0.383123622 0.017526147 0.020095216
HIP1 0.062598512 0.037507479 0.02014331
HLA.DPB1a 0.342394548 -0.019117511 0.02013547
HLA.DPB1b 0.122101699 -0.031139947 0.020142112
HLA.DPB1c 0.064544658 -0.037203842 0.020127614
HSD17B12 0.269026542 0.022243256 0.020124047
HSPA1A 0.433256989 -0.015733446 0.020077641
IRF4 0.980961258 -0.000475232 0.019914446
KCNN2 0.562593088 0.011575289 0.019992068
LDB1 0.462430681 0.014750708 0.020073086
LOC644145 0.850606075 0.003781261 0.020076229
MBP 0.456062468 0.014958583 0.020069316
MIA3 0.002200304 0.061630631 0.020128944
MRPL11 0.992977744 -0.00017814 0.020240506
MTHFD1 0.565546446 0.011500589 0.020014159
MYOM2 0.343226982 -0.018943889 0.019987045
NAV2 0.992285217 0.00019293 0.019953194
NETO2 0.138130776 0.029501295 0.019895826
NHSL1 0.913595924 0.002173091 0.020027781
NUDT12 0.400931634 -0.016787799 0.019986466
NXPH1 0.840757915 0.004028197 0.020048344
OXTR 0.229058416 0.024023348 0.019973049
PCDHGA4 0.190823705 0.026114059 0.019962639
PDE4C 0.351437649 0.018751882 0.020124272
PFDN1 0.096931444 0.03320262 0.020002598
PRKCQ 0.277179108 0.021676679 0.019947651
PTCHD3 0.361575109 -0.018196431 0.019944186
RAB3Ca 0.712338829 0.007406633 0.020087584
RAB3Cb 0.621374143 0.009929851 0.020104829
RABIF 0.159838577 0.028282623 0.020121173
RGS2 0.547686274 0.011935226 0.019851326
RNF216L 0.836113333 0.004177228 0.020192766
SARM1 0.124975135 0.030868802 0.020120169
SCYL1 0.813336194 0.004739042 0.020070105
SGEF 0.687859475 0.008097946 0.020156092
SLAIN1 0.528171702 0.012658068 0.020066715
SNRPN 0.887211913 0.00283658 0.019999444
SOCS3 0.940459415 -0.001503304 0.020126633
SORCS2 0.208087526 0.02523878 0.020049203
T 0.06099941 0.037496029 0.020013906
TBCCD1 0.74568301 -0.006500714 0.020043081
TBP 0.053654375 0.038586832 0.019997162
TMEM17 0.688208767 0.008048574 0.020056893
TRIM72 0.88554187 0.004250734 0.029529748
VENTX 0.13673696 0.029921666 0.020107929
XIST 0.737410414 -0.005439433 0.016223324
ZNF345 0.839358665 0.00404293 0.019944016
ZNF549 0.684729808 0.00814746 0.02006677
ZNF577 0.954753423 0.00113814 0.020059418
ZNF66 0.243316068 0.023459221 0.020106664
Overall none of these probes achieves Bonferoni significance. The only probe achieving nominal significance (p<0.002) after meta analysis of all 3 cohorts is MIA3, encoding melanoma inhibitory activity family, member 3. SNPs in this gene are associated with cardiovascular artery disease and myocardial infarction (Li X, et al Meta-analysis identifies robust association between SNP rs17465637 in MIA3 on chromosome 1q41 and coronary artery disease. Atherosclerosis. 2013 Nov;231(1):136-40.)
Birth weight genetic risk score analysis
THE ANALYSIS OF THE ASSOCIATION OF LOWER BIRTH WEIGHT RISK SCORE WITH LATER LIFE ADIPOSITY MEASURES:
Analysis done for cohorts: NFBC1966 (n=4912) and Hertfordshire Cohort Study (n=2397). A genetic risk score for lower birth weight was calculated using 17 birth weight SNPs which were available genotyped in HCS. The association of these individual SNPs with later life adiposity measures were also examined. The selected responses were BMI, weight, waist circumference and hip circumference.
Grs= genetic risk score for birthweight
RESPONSE: BMI NFBC1966 (n=4912) HCS (n=2397)
B SE p B SE p
grs -0.00225 0.00080 0.00476 0.00088 0.00112 0.43195
Betas, standard errors and p-values for the association with adult bmi. Adjusted for sex and age. P-value bold, if <0.05 after Bonferroni correction.
RESPONSE: WEIGHT NFBC1966 (n=4912) HCS (n=2397)
B SE p B SE p
grs -0.17879 0.06160 0.00372 0.05650 0.08752 0.51867
Betas, standard errors and p-values for the association with adult weight. Adjusted for sex, age and height. P-value bold, if <0.05 after Bonferroni correction.
RESPONSE: WAIST CIRCUMFERENCE NFBC1966 (n=4912) HCS (n=2397)
B SE p B SE p
grs -0.12727 0.05629 0.02381 0.04650 0.08235 0.57238
Betas, standard errors and p-values for the association with adult waist circumference. Adjusted for sex, age and height. P-value bold, if <0.05 after Bonferroni correction.
RESPONSE: HIP CIRCUMFERENCE NFBC1966 (n=4912) HCS (n=2397)
B SE p B SE p
grs -0.00131 0.00039 0.00085 0.00020 0.00059 0.73720
Betas, standard errors and p-values for the association with adult hip circumference. Adjusted for sex, age and height. P-value bold, if <0.05 after Bonferroni correction.
The analysis of the association of lower birth weight risk score with adult blood pressure:
Analysis done for: NFBC1966 (n=4259), using the same risk score as earlier.
Systolic Diastolic
Adjustments Beta 95 % CI Beta 95 % CI
sex -0.17 -0.3 -0.04 -0.08 -0.2 0.03
sex, birth vars -0.19 -0.32 -0.06 -0.09 -0.2 0.03
sex, birth vars, adult vars -0.14 -0.26 -0.02 -0.05 -0.17 0.06
sex, birth vars, adult vars, mother vars -0.15 -0.28 -0.03 -0.05 -0.16 0.06
Betas and their 95 % confidence intervals for the effect of birth weight risk score on systolic and diastolic blood pressures at 31 years with different adjustments. Birth vars: birth weight, gestational age, birth length. Adult variables: bmi at 31 years, smoking status, alcohol consumption. Mother variables: parity, family SES, mother's BMI.
References:
Das S.: Genetic and environmental correlates of growth patterns leading to obesity. A thesis based on Northern Finland Birth Cohort Data, completed in 2014 (Imperial College London/University of Oulu), under primary supervision of Professor Marjo-Riitta Jarvelin.
Das S et al: Common variants at LEPR and LEPROT, FTO, TFAP2B and GNPDA2 are associated with childhood growth trajectories, linking early growth with adulthood obesity and metabolism. Submitted, 2015.
Horikoshi M, .....Järvelin M-R, Timpson NJ, Prokopenko I, Freathy RM.: New loci associated with birth weight identify genetic links between intrauterine growth and adult height and metabolism. Nat Genet. 45, 76-82, 2013. doi:10.1038/ng.2477. Epub ahead of print. (IF 35.532)
Kaakinen M, Sovio U, Hartikainen A-L, Pouta A, Savolainen MJ, Herzig K- H, Elliott P, De Stavola B, Läärä E, Järvelin M-R.: Life course structural equation model of the effects of prenatal and postnatal growth on adult blood pressure. J Epidemiol Community Health 2014;68:1161–1167. doi:10.1136/jech-2013-203661. (IF 3.294).
Marttinen P, Pirinen M, Sarin AP, Gillberg J, Kettunen J, Surakka I, Kangas AJ, Soininen P, O'Reilly PF, Kaakinen M, Kähönen M, Lehtimäki T, Ala-Korpela M, Raitakari OT, Salomaa V, Järvelin M-R, Ripatti S, Kaski S.: Assessing multivariate gene-metabolome associations with rare variants using Bayesian reduced rank regression. Bioinformatics. 2014 Jul 15;30(14):2026-34. doi: 10.1093/bioinformatics/btu140. Epub 2014 Mar 24. (IF 5.323)
Menni, C. et al. Metabolomic markers reveal novel pathways of ageing and early development in human populations. International journal of epidemiology 42, 1111-1119, doi:10.1093/ije/dyt094 (2013).
Richmond RC. et al. Prenatal exposure to maternal smoking and offspring DNA methylation across the lifecourse: findings from the Avon Longitudinal Study of Parents and Children (ALSPAC). Hum Mol Genet. 2014 Dec 30. pii: ddu739. [Epub ahead of print]
Ronkainen J, Huusko T, Soininen R, Mondini E, Cinti F, Mäkelä KA, Kovalainen M, Herzig K-H, Järvelin M-R, Sebert S, Savolainen M, Salonurmi T.: Fat mass- and obesity-associated gene Fto affects the dietary response in mouse white adipose tissue. Nature Communications, in press 2015.
Vimaleswaran KS…..Järvelin M-R, Hyppönen E.: Association of vitamin D status with arterial blood pressure and hypertension risk: a mendelian randomisation study. Lancet Diabetes Endocrinol. 2014; 2: 719–29.
The association of birth measures and later life metabolomics:
The effect of birth measures on 25 selected metabolites were examined in NFBC1986 (n=2391), HCS (n not available) and TwinsUK (n not available). The metabolites were selected by King’s College London.
Meta-analysis: 12 metabolites significant after the meta-analysis with NFBC1986 (n=1170), HCS (n=1011) and TwinsUK (n=5036):
METABOLITE PValue_ Beta SE
Aspartate 0.0350 -0.0286 0.0136
Glycine 0.0001 0.0468 0.0122
g-Glutamylleucine 0.0003 -0.0476 0.0130
g-Glutamylvaline 0.0064 -0.0305 0.0112
Isovalerylcarnitine 0.0002 -0.0535 0.0142
Methylbutyroylcarnitine 7.08E-05 -0.0520 0.0131
Propionylcarnitine 0.0006 -0.0486 0.0141
Urea 4.63E-05 -0.0529 0.0130
Uric-acid 0.0002 -0.0547 0.0146
2-hydroxy-butyric-acid 2.05 E-06 -0.0625 0.0113
Glycosyltryptophan 0.0068 -0.0307 0.0113
3-4-hydroxyphenyllactate 0.0024 -0.0401 0.0132
A receiving operator characteristic curve was then performed to investigate how much were metabolites in adult life explained by low birth weight. ROC curves for birth weight metabolomic risk score in females from TwinsUK, NFBC86 and HCS. The response: low birth weight tertile. The explanatory variables: 11 metabolites listed in the meta-analysis table, excluding g-Glutamylvaline for severe collinearity, adjusted for bmi (see also table below).
These metabolites are also individually associated with blood pressure or insulin resistance in TwinsUK and HSC as the following table shows. A manuscript reporting these results is currently being drafted.
TwinsUK SBP DBP HOMA-IR
Beta[95%CI] P Beta[95%CI] P Beta[95%CI] P
2-hydroxy-butyric acid 0.59[0.08,1.09] 2.32E-02 0.22[-0.11,0.55] 2.32E-02 -0.04[-0.08,-0.01] 2.35E-02
3,4 -hydroxyphenyllactate 1.04[0.54,1.54] 5.19E-05 0.78[0.45,1.12] 5.19E-05 0.08[0.05,0.11] 6.51E-08
Aspartate 0.63[0.19,1.07] 5.55E-03 0.58[0.28,0.88] 5.55E-03 0.11[0.08,0.14] 2.58E-13
g-Glutamylleucine 0.36[-0.14,0.86] 1.59E-01 0.29[-0.05,0.63] 1.59E-01 0.1[0.07,0.14] 3.57E-08
g-Glutamylvaline 0.45[-0.04,0.94] 7.51E-02 0.49[0.16,0.82] 7.51E-02 0.15[0.11,0.18] 6.56E-15
glycine -0.24[-0.75,0.28] 3.65E-01 -0.32[-0.64,0] 3.65E-01 -0.08[-0.11,-0.05] 9.16E-08
glycosyltryptophan 0.25[-0.28,0.79] 3.54E-01 0.2[-0.16,0.56] 3.54E-01 0.01[-0.02,0.05] 4.63E-01
Isovalerylcarnitine -0.09[-0.57,0.39] 7.21E-01 0.06[-0.26,0.39] 7.21E-01 0.08[0.05,0.1] 1.86E-07
Methylbutyroylcarnitine 0.42[-0.07,0.92] 9.32E-02 0.39[0.06,0.72] 9.32E-02 0.11[0.07,0.14] 2.48E-11
Propionylcarnitine 0.09[-0.41,0.59] 7.30E-01 0.11[-0.22,0.44] 7.30E-01 0.09[0.06,0.12] 4.13E-09
urea 0.61[0.07,1.16] 2.79E-02 0.26[-0.08,0.6] 2.79E-02 0.04[0,0.07] 3.32E-02
uric Acid 1.09[0.57,1.6] 3.36E-05 0.68[0.34,1.02] 3.36E-05 0.1[0.07,0.13] 6.40E-11
HCS SBP DBP
metabolite Beta P Beta P
3,4-hydroxyphenyllactate 1.58[0.41,2.74] 8.32E-03 0.65[0.04,1.27] 3.75E-02
aspartate 0.42[-0.74,1.58] 4.78E-01 -0.62[-1.23,-0.01] 4.75E-02
g-glutamylleucine 0.18[-1.29,1.65] 8.09E-01 0.65[-0.14,1.43] 1.07E-01
g-glutamylvaline -0.77[-1.99,0.44] 2.14E-01 -0.13[-0.78,0.52] 6.98E-01
glycine 0.68[-0.54,1.9] 2.72E-01 0.25[-0.38,0.89] 4.33E-01
glycosyltryptophan 1.76[0.53,3] 5.24E-03 -0.01[-0.67,0.64] 9.65E-01
hydroxybutyric acid 0.81[-0.3,1.92] 1.52E-01 0.12[-0.49,0.73] 6.98E-01
isovalerylcarnitine -0.06[-1.25,1.13] 9.21E-01 0.06[-0.59,0.72] 8.51E-01
methylbutyroylcarnitine -0.16[-1.39,1.06] 7.95E-01 -0.04[-0.67,0.6] 9.07E-01
propionylcarnitine -0.72[-1.92,0.48] 2.41E-01 -1.1[-1.75,-0.46] 7.94E-04
urea 0.68[-0.49,1.84] 2.56E-01 0.08[-0.54,0.71] 7.94E-01
uric acid 1.27[0.17,2.36] 2.33E-02 0.99[0.43,1.55] 5.89E-04
Summary of WP6 :
1-The common pathways related to epigenetics and genetic associations with birthweight in twins highlight molecular pathways related adult height. The pathways related only to epigenetics are linked to various adult diseases including cardiovascular disease
2- Joint analyses of singletons and twins found consistency in associations with birthweight within singletons for all three technologies (genetics, epigenetics, metabolomics) but only for metabolomics between twins and singletons. Genetic results in singletons are consistent an a genetic risk score for birthweight correlates with measures of metabolic disease and blood pressure in adults A manuscript summarising these results is currently in preparation.
3- Epigenetic analyses resulted in only one gene being nominally (but not Bonferroni) significant between twins and singletons. This is the MIA3 gene which is known to be associated with cardiovascular disease
4- Metabolomics results in all three cohorts are consistent, and a metabolomic signature in adults is seen related to birthweight (AUC= 0.6). The metabolites associated with birthweight are also associated with blood pressure and insulin resistance in adults. A manuscript summarising these results is currently in preparation.
WP7- Data Integration and Health Benefit Evaluation
The main objective of Tasks 7.1-7.3 is to implement an integrated platform that will make available to its users an extensive database of the correlations between early life events and ageing outcomes linking datapoints to methylation, metabolomic, genetic and post-translational modification (PTM) results, allowing future comparison with data derived from other studies. In order to achieve this the following steps/objectives have to be completed:
1. Study the available data from the cohorts and identify what kind of information and datasets will be integrated into the database [DONE]
2. Define the User and Technical Requirements and Applications needed [DONE]
3. Design the System Architecture and Mock-ups of the applications’ GUI [DONE]
4. Develop, Integrate and test the applications of the system and database [DONE]
5. Define traits and then start using the platform and evaluate the results [DONE]
Summary: The main objective of this WP was to evaluate the impact of the identified biomarkers on health benefit and the possibility of their use within the context of the ageing European population and given the prevalence of the age-related diseases under study by EurHEALTHAgeing. In WP7, an integrated platform was implemented in order to help the researchers to compare the data derived from EurHealth studies with those that come from other studies based on the correlations between early life events and ageing outcomes linking datapoints to methylation, metabolomic, genetic and post-translational modification (PTM) results. All tasks have been completed as planned in the original workplan.
Task 7.1 (Requirements Analysis) has been completed and the corresponding deliverable has been submitted on time. The state of the art of existing protocols and data formats related to each partner’s dataset has been studied and used in order to define the correct data structures for the SW to be developed.
This input was used in Task 7.2 (Static and Dynamic MockUps) in order to design the Graphical User Interface (GUI) of the software and use this to collect the “users’ feedback”.
Task 7.3. (Technical Implementation, Testing and Deployment) is the actual EurHealthAgeing platform running on a web server. The users can access the database through a web browser and the credentials that have been provided to them. The corresponding deliverable D7.3 serves as the user manual of EurHealthAgeing Portal that contains all essential information for the user to make full use of the system.
In Task 7.4. (System Operation and Evaluation) the users completed the population of the database using the system provided services and applications.
The operation of the platform services was evaluated based on the evaluation methodology and criteria that were defined in the internal deliverable of T7.2.
Task 7.1. Requirements Analysis:
The main focus of this task was to study the available data from the cohorts and identify what kind of information and datasets will be integrated into the database thus providing an encompassing set of requirements for the structures and data that was used. Moreover all the valuable sources for the different domains involved in the project in accordance with the user requirements were identified and incorporated in our system.
Task 7.2 .Design the System Architecture and Design Mock-ups of the applications’ GUI
n this Task, the state of the art of existing protocols and data formats related to each partner’s dataset was studied. Furthermore we examined the existing technical state of the art and tried to manipulate them according to our platform’s needs. In this Task we defined the key building blocks of the architecture which were the basis for the technical development within T7.3. During T7.2 the system’s architecture and Database Schema were defined while the GUI was formed using static MockUps of the EurHealthAgeing Portal.
Task 7.3. Technical Implementation, Testing and Deployment
By using the available information and results driven from the previous WPs we developed our platform and the meta-database that stores the results of EurHealthAgeing studies. During this task, the implementation of the database schema, the web page development and the web server configuration took place in order to deliver the final platform to the users. The corresponding deliverable D7.3 includes a description of the system functions and capabilities, contingencies and alternate modes of operation, and step-by-step procedures for system access and use.
Task 7.4. System Operation and Evaluation
During this task users populated the database with their data and corresponding findings using the system provided services and applications. This was an opportunity for the users to actually perform a hands-on evaluation of the provided functionalities, ease-of-use and the other criteria as described in the evaluation criteria internal document. The outcome of this evaluation is summarized in D7.4 Evaluation Report
WP10- Ethics
During the length of the project no ethical issues have arisen. The external ethical advisor reviewed all the relevant documentation and found it satisfactory. She has participated at the project miterm and final reviews and assessed that for each workpackage and each participant institution proper ethical and data management protocols and procedures are in place and being performed. She has drafted a final report.
Potential Impact:
Impact
The EurHEALTHAgeing project has made a substantial contribution to the understanding of the early developmental molecular determinants that can affect healthy aging and furthered European scientific excellence in this area as follows:
1. It generated an international research network that can continue to foster scientific excellence in the field of determinants of healthy aging
2. Has increased our understanding of human biological variation across the lifespan in health and disease by studying genes and pathways.
3. Has identified molecular and genetic pathways involved in low birthweight that translate into higher risk of age –related diseases. This is particularly true for metabolomics markers.
4. A strong participation of SMEs in the projects has helped innovation in this area. European research intensive SMEs focussed on cutting edge methods of identifying early developmental markers of healthy aging have developed and optimised biomarker assays as part of this project
5. EurHEALTHAgeing has generated new knowledge to enable effective translation of biomarker research (an impact with broad applicability), specifically a better understanding of the strengths and weaknesses of various high-throughput genomic methods for biomarker identification. In particular the project has yielded important information and an understanding of the differences in markers between twins and singletons, and which technologies are applicable to both (metablomics, post translational collagen modifications) and which aren’t ( genetics, epigenetics).
6. EurHEALTHAgeing has also had positive impact on scientific research and technological development in Europe by publication of the research outcomes in peer-reviewed journals thereby raising the quality and enhancing the excellence of EU science by placing the EU at the cutting edge of this area of aging research.
7. The results from EurHEALTHAging have received worldwide coverage by the press, in particular the metabolomics work and have raised public awareness of the study of molecular pathways of aging in Europe.
Main dissemination activities and exploitation
The main objectives for dissemination of the project were:
1. To communicate to the public, the EU and the wider scientific and health-care community the novel genes and pathways discovered by EurHEALTHAgeing and their potential values as diagnostics
2. To encourage interest in ageing science through mentoring, workshops and summer schools
3. To communicate and collaborate with SMEs and other industrial partners to translate the results into clinical application
4. To liaise and work closely with the other EU funded age-research programmes
5. To disseminate to the widest audience of the general public to promote greater public engagement and dialogue between the scientists and society.
A dedicated public website was set up at the beginning of the project, which also contains a password protected area for members of the Consortium to share information. There are dedicated pages to present the project to the wider public, and to provide links to other ageing related research.
Period 1
Publications
Valdes, AM, Glass D, Spector T, OMICS technologies and the study of human ageing Nature Reviews Genetics, 14, Sep 2013, 692-702
Menni, C et al Metabolomic markers reveal novel pathways of ageing and early development in human populations Int J Epidemiology, July 8 2013; 1-9
Menni C et al Targeted metabolomics profiles are strongly correlated with nutritional patterns in women Metabolomics 2013 Apr:9(2) 506-514
PhD thesis: Kaakinen, Marika (2013) Genetic and life course determinants of cardiovascular risk factors: structural equation modelling of complex relations. Acta Universitatis Ouluensis. Series D, Medica 1193, 2013.
The thesis includes the following published articles:
Kaakinen M, Läärä E, Pouta A, Hartikainen AL, Laitinen J, Tammelin TH, Herzig KH, Sovio U, Bennett AJ, Peltonen L, McCarthy MI, Elliott P, De Stavola B, Järvelin MR (2010) Life-course analysis of a fat mass and obesity-associated (FTO) gene variant and body mass index in the Northern Finland Birth Cohort 1966 using structural equation modeling. Am J Epidemiol 172(6): 653-65
Ducci F*, Kaakinen M*, Pouta A, Hartikainen AL, Veijola J, Isohanni M, Charoen P, Coin L, Hoggart C, Ekelund J, Peltonen L, Freimer N, Elliott P, Schumann G*, Järvelin MR* (2011) TTC12-ANKK1-DRD2 and CHRNA5-CHRNA3-CHRNB4 influence different pathways leading to smoking behavior from adolescence to mid-adulthood. Biol Psychiatry 69(7): 650-60.
Kaakinen M, Ducci F, Sillanpää MJ, Läärä E, Järvelin MR (2012) Associations between variation in CHRNA5-CHRNA3-CHRNB4, body mass index and blood pressure in the Northern Finland Birth Cohort 1966. PLoS One 7(9): e46557.
Selected Conferences (full list uploaded to portal)
A. Weinhäusel, R. Kallmeyer, R.V. Pandey, I. Visne, E. Dilaveroglu, A. Yildiz, and A. Kriegner (2012) “MutAid”- A comprehensive sequence variance analyses tool for pipelined human molecular genetic diagnostics. Bio-IT World Europe, Vienna, October 9-11, 2012.
A. Weinhäusel, (2012) Results from Serum-Tumorautoantibody profiling of Breast, Colon, Lung and Prostate Cancers using a 16k protein-µ-array for improving minimal invasive diagnostics. Systems Biology Europe 2012, Madrid, Spain, October 16-17, 2012
A. Weinhäusel, M. Wielscher, W. Pulverer, M. Sonntagbauer, A. Kriegner, K. Vierlinger, M. Hofner, , and C. Nöhammer (2012) DNA methylation cancer-biomarker validation using restriction enzyme-based qPCR application Application of high throughput qPCR in DNA-methylation biomarker research. Fluidigm Seminar, Frankfurt, Oct. 25, 2012
Weinhäusel A, (2012). Epigenetic Markers in Human Cancer. 36th Seminar of the Austrian Society on Surgical Research, “Bench to Bedside” in Surgical Oncology, Vienna, Nov 23-24th, 2012
A. Weinhäusel, R. Kallmeyer, R.V. Pandey, I. Visne, E. Dilaveroglu, A. Yildiz, and A. Kriegner (2013) “MutAid”- A comprehensive sequence variance analyses tool for pipelined human molecular genetic diagnostics. European Lab Automation 2013, Next-Gen Sequencing, June 6-7, 2013, Hamburg, Germany.
Weinhäusel A, Wielscher M, Reithuber E, Sonntagbauer M, Hofer P, Pulverer W, Pichler R, Kriegner A, Liloglou T, Vierlinger K, and Gsur A (2013). DNA methylation biomarkers - from tissue to body fluids. Symposium of the Bavarian Red Cross Biobank: Epigenetic Biomarkers & Complex Diseases”, June 28, 2013, Schloss Höhenried, Germany.
Jarevelin Marjo-Ritta, Genetic and early life determinants of health and diseases – current evidence from humans and animals, II Forum Discussion, Lipids and Programming, Granada, Spain
A.Weinhäusel, M. Wielscher, E. Reithuber, M. Sonntagbauer, P. Hofer, W. Pulverer, R. Pichler, A. Kriegner, T. Liloglou, K. Vierlinger and A. Gsur (2013). DNA-Methylation Testing/Biomarker Validation Using High Throughput qPCR: qPCR & Digital PCR Congress, September 9-10, 2013, Lyon, France
Posters
Hofner M., Fürst F., Kriegner A., Gruber G., Weinhäusel A. (2012) Altered DNA methylation patterns in osteoarthritis. Tissue Remodeling in Ageing and Disease - Emerging Insights into a Complex Pathology.”, Vienna, Austria, 27- 28 March 2012.
Weinhaeusel A., Wielscher M., Sonntagbauer M., Pulverer W., Hofner M., Kriegner A., Vierlinger K., and Noehammer C. (2012) High throughput qPCR DNA methylation marker testing and validation. European Society of Human Genetics Conference, Nürnberg, Germany, June 23-26, 2012.
Kallmeyer R (2013a): Development of pipelines for high-throughput data analyses, primer and assay design. Functional Genomics and Proteomics - Applications. Frankfurt, Germany, 31.01.2013.
Kallmeyer R (2013b): High Throughput Primer and Assay Design Pipelines for Epigenetics. Jahrestagung der Gesellschaft für Humangenetik. Dresden, Germany, 20.03.2013.
Kaakinen M, Polymorphisms in CHRNA3-CHRNA5-CHRNB4 are associated with body mass index and systolic blood pressure in smokers in the Northern Finland Birth Cohort 1966 62nd annual meeting of the American Society of Human Genetics, San Francisco, USA, November 2012
Jarvelin MarjoRitta Variants at LEPR, FTO, GNPDA2 and TFAP2B genes are associated with early adiposity phenotypes, 62nd annual meeting of the American Society of Human Genetics, San Francisco, USA, November 2012
Technology training (Task 9.3)
design tools are implemented in XworX platform (www.xworx.org)
training courses are established
manuals available online (www.xworx.org/#!manual/cnvu)
video tutorial available online (www.xworx.org/#!tutorial/cspb)
webinar established pipelines will be made available in Q1/2014
YouTube videos for each “Design Pipeline workflow” will be added
Studentships (Task9.7)
It was planned that within the EurHealthAging project 6 diploma/master students will be trained in software engineering and bioinformatics as well as in molecular biology laboratory work. A major improvement to the initial proposal is that the diploma/master students working in the lab will not only be trained in epigenetic and SNP-assay development, but also in the development of assays for bisulfite sequencing PCR (BSP).
Up to now 10 people have been working within the EurHealthAging project. One of them being a Masterstudent (Beikircher G), two of them working on their bachelor thesis (Ziegler S, Orhan N) and seven of them joined AIT for an internship (Mikulasek B, Breitenberger C, Buck A, Mayer T, Wintersperger S, Eder F, Roffeis S). The attached table contains information on the students and their activities at AIT. That number of students will still raise and is therefore also an improvement compared to the initial proposal. This is made possible by the Austrian Research Promotion Agency (FFG) which offers financial support for students working in research facilities.
Period 2
Partners have taken part in national and international conferences, and in events geared to the public. A Summer School was held during the period, which was open to a wider audience.
Summer School Programme
09.30 Opening of School: Ana Valdes and Tim Spector
09.40 Session I: Can Life Course Studies unravel age-related disease?
Session chairs: Ana Valdes & Tim Spector
09.40 Plenary Session-
Diana Kuh (Medical Research Centre)-
Healthy Aging across the Life Course
10.20 Discussion
10.30 Marika Kaakinen (University of Oulu) - Measuring Life Course
10.50 Discussion
11.00 Rachel Freathy (University of Exeter)
Early Growth Genetics- EGG Consortium
11.20 Discussion
11.30 Break
12.00 Session II: Linking Early Growth with aging via omics
Session chairs: Rachel Freathy and Marjo-Riita Jarvelin
12.00 Ana Valdes (University of Nottingham)-
Integrating –omics to investigate life course in twins
12.20 Discussion
12.30 Michelle Beaumont (Kings College London) - Microbiome Impact on Early Life
12.50 Discussion
13.00 - 14.00 Lunch break
14.00 Session III: Epigenetic studies shed light on Life course
Session chairs: Tim Spector and Marika Kaakinen
14.00 Hannah Elliot (University of Bristol)-The ARIES project
14.20 Discussion
14.30 Marie Loh (Imperial College London)-
Analysing epigenetics for complex traits
14.50 Discussion
15.00 Coffee break
-15.30 Jordana Bell (Kings College London)-
Epigenetic Studies of Twins
15.50 Discussion
16.00 Marjo-Riita Jarvelin (Imperial College London) -
The benefits of studying Life Course
16.30 Discussion of Life Course Projects
16.45 – 17.00 Ending of the School
The Summer School was attended by 50 people.
Publications
Numerous papers have been published in peer reviewed journals (full list in portal);
Menni C, Mangino M, Cecelja M, Psatha M, Brosnan MJ, Trimmer J, Mohney RP, Chowienczyk P, Padmanabhan S, Spector TD, Valdes AM. Metabolomic study of carotid-femoral pulse-wave velocity in women. J Hypertens. 2014 Dec 8. [Epub ahead of print]
Menni C, Kiddle SJ, Mangino M, Viñuela A, Psatha M, Steves C, Sattlecker M, Buil A, Newhouse S, Nelson S, Williams S, Voyle N, Soininen H, Kloszewska I, Mecocci P, Tsolaki M, Vellas B, Lovestone S, Spector TD, Dobson R, Valdes AM. Circulating Proteomic Signatures of Chronological Age.
J Gerontol A Biol Sci Med Sci. 2014 Aug 14. pii: glu121. [Epub ahead of print]
Metrustry SJ, Edwards MH, Medland SE, Holloway JW, Montgomery GW, Martin NG, Spector TD, Cooper C, Valdes AM. Variants close to NTRK2 gene are associated with birth weight in female twins. Twin Res Hum Genet. 2014 Aug;17(4):254-61
Menni C, Keser T, Mangino M, Bell JT, Erte I, Akmačić I, Vučković F, Pučić Baković M, Gornik O, McCarthy MI, Zoldoš V, Spector TD, Lauc G, Valdes AM. Glycosylation of immunoglobulin g: role of genetic and epigenetic influences. PLoS One. 2013 Dec 6;8(12):e82558.
Serological identification of fast progressors of structural damage with rheumatoid arthritis. Siebuhr AS, Bay-Jensen AC, Leeming DJ, Platt A, Byrjalsen I, Christiansen C, van der Heijde D, Karsdal M. Arthritis Res Ther. 2013 Aug 14;15(4):R86. [Epub ahead of print]
Quantification of "end products" of tissue destruction in inflammation may reflect convergence of cytokine and signaling pathways - implications for modern clinical chemistry. Karsdal MA, Bay-Jensen AC, Leeming DJ, Henriksen K, Christiansen C. Biomarkers. 2013 Aug;18(5):375-8. doi: 10.3109/1354750X.2013.789084. Epub 2013 May 31.
Circulating protein fragments of cartilage and connective tissue degradation are diagnostic and prognostic markers of rheumatoid arthritis and ankylosing spondylitis. Bay-Jensen AC, Wichuk S, Byrjalsen I, Leeming DJ, Morency N, Christiansen C, Karsdal MA, Maksymowych WP. PLoS One. 2013;8(1):e54504. doi: 10.1371/journal.pone.0054504. Epub 2013 Jan 24.
Circulating citrullinated vimentin fragments reflect disease burden in ankylosing spondylitis and have prognostic capacity for radiographic progression. Bay-Jensen AC, Karsdal MA, Vassiliadis E, Wichuk S, Marcher-Mikkelsen K, Lories R, Christiansen C, Maksymowych WP. Arthritis Rheum. 2013 Apr;65(4):972-80. doi: 10.1002/art.37843
Taylor AE, Fluharty ME, Bjørngaard JH, Gabrielsen ME, Skorpen F, Marioni RE, Campbell A, Engmann J, Mirza SS, Loukola A, Laatikainen T, Partonen T, Kaakinen M, Ducci F, Cavadino A, Husemoen LL, Ahluwalia TS, Jacobsen RK, Skaaby T, Ebstrup JF, Mortensen EL, Minica CC, Vink JM, Willemsen G, Marques-Vidal P, Dale CE, Amuzu A, Lennon LT, Lahti J, Palotie A, Räikkönen K, Wong A, Paternoster L, Wong AP, Horwood LJ, Murphy M, Johnstone EC, Kennedy MA, Pausova Z, Paus T, Ben-Shlomo Y, Nohr EA, Kuh D, Kivimaki M, Eriksson JG, Morris RW, Casas JP, Preisig M, Boomsma DI, Linneberg A, Power C, Hyppönen E, Veijola J, Järvelin M-R, Korhonen T, Tiemeier H, Kumari M, Porteous DJ, Hayward C, Romundstad PR, Smith GD, Munafò MR.: Investigating the possible causal association of smoking with depression and anxiety using Mendelian randomisation meta-analysis: the CARTA consortium. BMJ Open. 2014 Oct 7;4(10):e006141. doi: 10.1136/bmjopen-2014-006141. IF 2.063
Marttinen P, Pirinen M, Sarin AP, Gillberg J, Kettunen J, Surakka I, Kangas AJ, Soininen P, O'Reilly PF, Kaakinen M, Kähönen M, Lehtimäki T, Ala-Korpela M, Raitakari OT, Salomaa V, Järvelin M-R, Ripatti S, Kaski S.: Assessing multivariate gene-metabolome associations with rare variants using Bayesian reduced rank regression. Bioinformatics. 2014 Jul 15;30(14):2026-34. doi: 10.1093/bioinformatics/btu140. Epub 2014 Mar 24. (IF 5.323)
Kaakinen M, Sovio U, Hartikainen A-L, Pouta A, Savolainen MJ, Herzig K- H, Elliott P, De Stavola B, Läärä E, Järvelin M-R.: Life course structural equation model of the effects of prenatal and postnatal growth on adult blood pressure. J Epidemiol Community Health 2014;68:1161–1167. doi:10.1136/jech-2013-203661. (IF 3.294).
Vimaleswaran KS, Cavadino A, Berry DJ; LifeLines Cohort Study investigators, Jorde R, Dieffenbach AK, Lu C, Alves AC, Heerspink HJ, Tikkanen E, Eriksson J, Wong A, Mangino M, Jablonski KA, Nolte IM, Houston DK, Ahluwalia TS, van der Most PJ, Pasko D, Zgaga L, Thiering E, Vitart V, Fraser RM, Huffman JE, de Boer RA, Schöttker B, Saum KU, McCarthy MI, Dupuis J, Herzig KH, Sebert S, Pouta A, Laitinen J, Kleber ME, Navis G, Lorentzon M, Jameson K, Arden N, Cooper JA, Acharya J, Hardy R, Raitakari O, Ripatti S, Billings LK, Lahti J, Osmond C, Penninx BW, Rejnmark L, Lohman KK, Paternoster L, Stolk RP, Hernandez DG, Byberg L, Hagström E, Melhus H, Ingelsson E, Mellström D, Ljunggren O, Tzoulaki I, McLachlan S, Theodoratou E, Tiesler CM, Jula A, Navarro P, Wright AF, Polasek O; International Consortium for Blood Pressure (ICBP); Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortium; Global Blood Pressure Genetics (Global BPGen) consortium; Caroline Hayward, Wilson JF, Rudan I, Salomaa V, Heinrich J, Campbell H, Price JF, Karlsson M, Lind L, Michaëlsson K, Bandinelli S, Frayling TM, Hartman CA, Sørensen TI, Kritchevsky SB, Langdahl BL, Eriksson JG, Florez JC, Spector TD, Lehtimäki T, Kuh D, Humphries SE, Cooper C, Ohlsson C, März W, de Borst MH, Kumari M, Kivimaki M, Wang TJ, Power C, Brenner H, Grimnes G, van der Harst P, Snieder H, Hingorani AD, Pilz S, Whittaker JC, Järvelin M-R, Hyppönen E.: Association of vitamin D status with arterial blood pressure and hypertension risk: a mendelian randomisation study. Lancet Diabetes Endocrinol. 2014; 2: 719–29.
Das S.: Genetic and environmental correlates of growth patterns leading to obesity. A thesis based on Northern Finland Birth Cohort Data, completed and published in 2014 (Imperial College London/University of Oulu), under primary supervision of Professor Marjo-Riitta Jarvelin.
Ronkainen J, Huusko T, Soininen R, Mondini E, Cinti F, Mäkelä KA, Kovalainen M, Herzig K-H, Järvelin M-R, Sebert S, Savolainen M, Salonurmi T.: Fat mass- and obesity-associated gene Fto affects the dietary response in mouse white adipose tissue. Nature Communications, in press 2015.
Das S et al: Common variants at LEPR and LEPROT, FTO, TFAP2B and GNPDA2 are associated with childhood growth trajectories, linking early growth with adulthood obesity and metabolism. Submitted, 2015.
Conferences & presentations
Pulverer W (2014): Phenotype associated methylation differences in Monozygotic Twins – a genome wide approach. Jahrestagung der Gesellschaft für Humangenetik. Essen, Germany, Mar. 19-21, 2014. Poster
Kallmeyer R (2014a): Combined High Throughput Assay Design and Analysis Pipelines for DNA methylation analyses. Jahrestagung der Gesellschaft für Humangenetik. Essen, Germany, Mar. 19-21, 2014. Poster
Kallmeyer R (2014b): Infinium Methylation Assay Analyses Pipeline – A Key for genome-wide DNA methylation analyses. Jahrestagung der Gesellschaft für Humangenetik. Essen, Germany, Mar. 19-21, 2014. Poster
Kallmeyer R (2014c): Novel high throughput analysis software pipelines for improving DNA methylation analysis. Austrian Association of Molecular Life Sciences and Biotechnology 6th Annual meeting. Vienna, Austria, Sept. 15-18, 2014. Poster
Weinhäusel, M. Wielscher, E. Reithuber, M. Sonntagbauer, P. Hofer, W. Pulverer, R. Pichler, A. Kriegner, T. Liloglou, K. Vierlinger and A. Gsur (2013). DNA-Methylation Testing/Biomarker Validation Using High Throughput qPCR: qPCR & Digital PCR Congress, Lyon, France, Sept. 9-10, 2013. Talk.
W. Pulverer, A. Hessenberger, O. Koperek, K. Kaserer, A. Weinhäusel and K. Vierlinger (2014). Exploring the methylome of thyroid cancer at single C resolution: From screening to clinical diagnostics. AIT, 2nd Austrian Biomarker Symposium 2014, Vienna, Austria, Mar. 31, 2014. Talk
Weinhäusel A., Reithuber E., Sonntagbauer M., Wielscher M., Hofer P., Pulverer W., Nöhammer C., Kriegner A., Liloglou T., Hajduch M., Vierlinger K., and Gsur A.(2014). Qualification of Lung Cancer DNA methylation markers for liquid biopsy testing. 25. Jahrestagung der Deutschen Gesellschaft für Humangenetik , March 19. – 21, 2014, Essen, Germany
Weinhäusel A., et al. (2014) Immuno-proteomics based early cancer diagnostics - from antigenic proteins to antigenic peptides. 2nd Austrian Biomarker Symposium 2014, March 31 - April 1, 2014, Vienna, Austria. Talk.
Noehammer C., Wielscher M., Fuchs-Luna J, Gyurjan I., Hofner M., Kegler U., Stoeger L., Singer C., Längle F., Hofbauer J., Gsur A., Ziesche R., Vierlinger K. and Weinhaeusel A. (2014). DNA-methylation - and - autoantibodies based cancer diagnosis from body fluids. Translational Medicine (EUSTM-2014), Sep 22 - 25 Sep, 2014 - Vienna, Austria
Andreas Weinhäusel, Istvan Gyurjan, Johana Fuchs-Luna, Regina Linhart, Gordana Wozniak-Knopp, Florian Rüker, Nina Pecha, Robert Zeillinger , Christine Rappaport, Christian Singer, and Johannes Söllner (2014). Immunopoteomics: protein and peptide microarrays for defining serum-IgG signatures for cancer diagnostics. 19th World Congress on Advances in Oncology and 17th International Symposium on Molecular Medicine, 9-11 October, 2014, Athens, Greece. Talk
Andreas Weinhäusel, Matthias Wielscher, Walter Pulverer, Albert Kriegner, Rainer Kallmeyer, Stefan Pabinger, Istvan Gyurjan, Johana Luna, Gernot Leeb, Karl Mach, Friedrich Längle, Johann Hofbauer, Christian Singer, Andrea Gsur, Johannes Söllner, Klemens Vierlinger , Christa Noehammer, Uwe von Ahsen and Martin Weber (2014). DNA-methylation- and autoantibody- biomarker development strategies for minimal invasive diagnostics. 3rd International Conference on Translational Medicine, November 3-5, 2014, Las Vegas, USA. Talk
Andreas Weinhäusel, Matthias Wielscher, Johana Luna , Istvan Gyurjan, Walter Pulverer, Klemens Vierlinger, Christa Noehammer, Johannes Söllner and Albert Kriegner (2014). DNA-methylation- and autoantibody- biomarker development strategies for minimal invasive diagnostics. 4th Munich Biomarker Conference, November 25-26, 2014, Munich, Germany. Talk
Anastasios Tagaris, Athanasios Anastasiou, Yannis Makris, Keith Grimaldi, Dimitrios Koutsouris
2014 An Extensive Database of the Correlations between Early Life Events and Ageing Outcomes, Medical Informatics Conference - MIE2014 - in Istanbul, August 31st – September 3rd
Series: Studies in Health Technology and Informatics
EBook: Volume 205: e-Health – For Continuity of Care
Pages: 1256 – 1256
Those peculiar Finns - did they really come from Asia? Marjo-Ritta Jarvelin, Finnish Institute, London 21st January 2014. Invited talk to the Anglo-Finnish Society
Liggins Institute, University of Auckland, Genetic Determinants of Early Growth and Development, Marjo-Ritta Jarvelin, (OULU) 28 February 2014. Talk.
Tamaki Campus, University of Auckland - Genetic Determinants of Early Growth and Development Marjo-Ritta Jarvelin (OULU) 3 March 2014. Talk
MRC Unit for Lifelong Health and Ageing at UCL - Lifecourse Studies in the Northern Finland Lifecourse Studies Programme, Marjo-Ritta Jarvelin, 16 April 2014. Talk
Eurosymposium on Healthy Ageing
Omic markers of aging using twin studies, Ana Valdes (KCL)
1 October 2014. Talk
List of Websites:
www.eurhealth.org
