Skip to main content

Genetic study of Common Hereditary Bowel Cancers in Hispania and the Americas

Final Report Summary - CHIBCHA (Genetic study of Common Hereditary Bowel Cancers in Hispania and the Americas)

Executive Summary:

The project has achieved the following.

1. Sample collection and preparation has been completed in Colombia, Mexico, Brazil and Portugal (Partners 1, 2, 3, 4, 9). This is a major component of the project on which all other components depend. It has been achieved with considerable effort, sometimes in difficult local environments. DNA has now been supplied and quality control ensured.

2. A new Latin American ancestry informative marker (AIM) panel has been developed by Partner 7 as part of a collaboration with groups in the USA.

3. Samples of over 200 Native Americans have been collected from Colombia (Partner 2), and many of these have been genotyped and analysed using the AIM panel.

4. Work is in progress to test the known colorectal cancer predisposition genes and SNPs in Latin America (Partners 2, 3, 4)

5. Mexican and Colombian samples (N>4,000) have been genotyped for about 1.2M SNPs comprising genome-wide tagging and custom panels. Initial quality control procedures and association analyses have been performed. Further analysis and meta-analysis are in progress.

6. European samples have been genotyped for uncommon coding variants using the Illumina Exome Array. The array content is mirrored on the arrays used for genotyping the Latin American samples and a combined analysis will be performed.

7. A method for allele-specific copy number assessment using SNP typing has been developed.

8. Consortium meetings with guest scientific speakers have been held in the UK, Mexico and Spain and USA, the latest being in June 2013 (all Partners).

10. Secondment of researchers from Colombia and Uruguay to European Partners for training and knowledge transfer has occurred.

It is important to note that the final work for the project will overrun the proposed the end date of 31st June 2013. There are several reasons for this that have accumulated over time, and these are detailed below. Critically, all samples and the great majority of the molecular data have been obtained. The molecular data awaited are from the validation phase of the GWAS and AMS. The validation phase is already funded from the CHIBCHA award and all samples are in place for this. The effort required to bring the project to fruition will almost entirely comprise data analysis, and this will be performed by existing staff from the Partners’ groups (principally Partners 1 and 7, with assistance from a new group set up by the ex-Programme Co-Ordinator (Dr Carvajal-Carmona) in UC Davis. No additional EU monies will be required for this work. We have rejected alternative strategies, such as analysing smaller numbers of samples, since this would jeopardise the scientific results from the CHIBCHA project.

Project Context and Objectives:

Colorectal cancer (CRC) is common in both sexes, has relatively poor outcome and has no major avoidable risk factor. Recent studies have shown that common inherited single nucleotide polymorphisms (SNPs) can increase cancer risk. We have shown CRC risk to be associated with SNPs on chromosomes 1q41, 3q26.2 8q23.3 8q24.21 10p14, 11q23.1 12q13.13 14q22.2 15q13.3 16q22.1 18q21.1 19q13.11 20p12.3 and 20q13.33. These variants account for <5% of the genetic risk of CRC, but will be very important when their effects are added to those of other, as-yet undetected CRC SNPs. A few genome-wide association studies (GWASs) based on populations of European descent are trying to identify the remaining common CRC genes. Evidence suggests that these studies will not be large enough on their own to detect all CRC SNPs, as: relative risks associated with most SNPs are modest; some disease alleles are rare, at least in Europe; and many variants may lie outside conventional gene boundaries or haplotype blocks. The admixed population from Latin American (LA) provides an exciting opportunity to identify new CRC genes that are more tractable to detection in LA, or have been missed by chance in European studies. We shall perform a detailed characterisation of Native American genomes from LA, use these to undertake an admixture mapping study, and perform a GWAS analysis on this same data set. We aim primarily to detect SNPs with effects in both LA and Europe, but also SNPs with effects specific to LA. Eventually, we aim to develop a polymorphism panel for predicting the risk of CRC in the general population, so that those at increased risk can be offered effective measures to prevent cancer. CRC is increasing in frequency in LA and prognosis is poorer than in Europe. We shall use our project as a focus for education about CRC, especially in LA. The study will also provide training for young LA researchers. Our work will provide a direct benefit to medical science and the populations of LA and Europe.

Project Results:

Core of the report for the period months 1-48: project objectives, work progress and achievements, project management

WP1, WP8. Management, dissemination, training and education (up to Month 48)

Main management duties and work toward achieving deliverables

Until leaving in late 2012, the Project Manager (L. Carvajal-Carmona) carried out all the managerial tasks and duties that focused on fulfilling the Objectives of the Management Work Package (WP1). Such duties included the co-ordination of all Consortium activities; the monitoring of the Scientific Directions that were set out by the Scientific Board; the constant monitoring of the results and progress of the scientific, training, managerial and dissemination aspects of the Consortium; the help in resolving Consortium disputes; the optimization of the data and sample flows and the timely reporting to the EU. All these activities had the support from the Project Co-ordinator, the Wellcome Trust Centre for Human Genetics Administration office, the University of Oxford European Office and the Partners in the Consortium.

The first successfully completed deliverables in WP1 were the setting up of the Consortium Agreement (Deliverable D1.1) and the approval of research protocol by local ethics committees (Deliverable D.1.2). The Consortium agreement was successfully negotiated and signed by all Partners on May 28th 2009. All necessary ethical approvals were obtained by May 2009 and copies of such approvals were sent to the EU office.

Another managerial task carried out by the Project Manager was setting up of timelines for patient and control sample collection and delivery of DNA samples for genotyping. These timelines were discussed and agreed during the first and second Consortium meetings. Overall, most of the managerial effort during the first 24 months of the Consortium focused on closely following the sample collection efforts carried out by the Partners in Portugal, Mexico, Colombia and Brazil. In the last phase of the study, the Project Manager dedicated most of his efforts to fulfil the other Deliverable in WP1, which consisted in preparing and submitting the 18-month, 36-month and final financial and scientific reports to the EU (Deliverable D1.3). During the final months of the study, the Project Manager dedicated his efforts to the coordination of sample preparation for the genotyping with SNPs arrays by Partner 7 (USC), the organization and transfer of the basic clinic-pathological data for all the genotyped samples and the co-organization of the last consortium meeting in Sacramento, during June 2013. The Project Manager was employed by the Consortium until October 2012 and has taken a position as Assistant Professor at the University of California, Davis. He however continued with his involvement in the study, including providing support in data analyses, training CHIBCHA funded Colombian and Mexican students in his group and co-organizing the last Consortium meeting in Sacramento.

Co-ordination of training activities

The Project Manager worked with Partners 1, 2, 3, 7 and 10 to co-ordinate training activities with young European and Latin American Scientist in the UK and Spain. Partner 7 sent Mrs Ceres Fernandez, a PhD student, for training in statistical genetic analyses to Partner 1’s laboratory in Oxford. Mrs Fernandez subsequently produced several quality publications using the knowledge and skills obtained during her training in Oxford and then successfully competed for an EU Marie Curie Postdoctoral Fellowship. She is currently working in Oxford on a project that involves the feasibility of using bowel cancer risk polymorphisms to stratify patients for bowel cancer screening. Partner 2 sent Mr Angel Criollo, a Colombian MSc student, and Dr Mabel Bohorquez, a Colombian clinical PhD student, for training in Oxford. Mr Criollo completed his Masters degree and received the highest distinction by the University of Tolima for the quality of his CHIBCHA based research study. Mr Criollo went further to win a PhD studentship from the Colombian Science Foundation (Colciencias) and is currently carrying out his PhD with Partner 7. Dr Bohorquez continues with her PhD and is currently finishing CRC somatic genetic analyses for her PhD thesis and publication. Partner 3 sent two Mexican PhD students for training at USC (Partner 7) in SNP genotyping and Partner 9 sent her PhD student (Ms Valentina Colistro) for training in statistical genetic analyses in Oxford. Finally, the Project Manager, himself a Latin American scientist, also benefited from his participation in CHIBCHA. He is currently an Assistant Professor at University of California (USA), where has continued his collaborations with the various Partners to ensure that the unique resources generated by the CHIBCHA collaborations continue to be used for a better understanding of CRC genetics and epidemiology in the region. Overall, one European and six Latin American postgraduate students carried out training visits during the duration of the study. At every Consortium meeting, these and several other students, also received informal training and advise from senior Partners in the study.

Major organisational challenges during the course of the study

A major challenge in the study was the initial phase of sample collection in Latin America. A timely collection of samples was crucial for the proposed project timetable because such samples were needed for the experiments proposed in WP 3-7. The first problems were experienced with Ecuador (an initial applicant who subsequently failed to respond to any communication and did not sign the consortium agreement), Argentina (who subsequently left the Consortium having performed no work for it) and Mexico. Early on in the study, continual close monitoring of the progress in Mexico resulted in improved recruitment in this country. The Project Manager worked closely with this latter Partner in obtaining permission, from the National Mexican Government, to export the Mexican DNA samples to Europe. This was a major difficulty during this reporting period since a change in legislation in Mexico required Partner 3 to obtain permission to export the samples. The project manager sent to Partner 3 all the necessary documents and information required to process and obtain their exporting license, which was obtained in April 2012.

Delays in sample provision have continued to dog the project (see Deliverables 2.1 and 2.2 below). Our view throughout has been that it is essential for the project’s success to obtain sufficient samples and to provide the means for them to be genotyped. Thus, whilst the delays have been frustrating, we have agreed on two priorities by the end of the 48-month project period, (i) supply of sufficient blood and DNA samples to empower the project and (ii) provision of genotyping infrastructure and consumables to allow those samples to be genotyped as per the original project plan. These have been achieved within budget. Our contingency plan has been to delay the analysis phase of the project. It was always planned largely to use existing staff for this phase of the project and we confirm that this will be done by Partner 1, Partner 10 and the ex-Project Manager Dr Carvajal-Carmona’s new team at UC Davis.

Disputes during the study

During the 42 months of the study, only two minor disputes existed within the Consortium. The first dispute was with the Argentinian Partner, FUDEIG. Argentina, an early recruiting Partner in the study, officially left the Consortium due to lack of progress after the Consortium decided to ask them to leave during the Second Consortium meeting in Cancun.

The second issue was with KBIOSCIENCE. KBIOSCIENCE, the SME Partner which left the Consortium by mutual agreement without any funds having been transferred. Owing to changes in the SNP typing market, KBioscience wished to focus on their existing commercial activities rather than to take part in a collaborative project.

Overall, the departure of these two Partners did not affect the Consortium’s overall plan, since the resources allocated to them, for sample collection and genotyping were re-deployed to the other Partners in Consortium. All these changes were formally be proposed, and accepted by the EU office, in revised Technical Annex documents. However, the approval of the revised Technical Annex only occurred late in 2012 and hence monies could not be provided to other Partners for some time. Overall, the absence of FUDEIG has contributed significantly to the project’s sample collection delays and timescale, but not to the overall chances of scientific success.

Consortium meetings

The Project Manager organised or co-organised, in collaboration with Local Partners, very successful Consortium meetings. These meetings were extremely important because they allowed the Project Manager Office and the Consortium to monitor progress, to resolve disputes and to make changes to the study that will help adapt to existing challenges. Such meeting always had both business and scientific components. In the latter, external experts were invited to present in topics that were highly related to the aims of the study.

The first meeting of the Consortium took place in Oxford, at the Wellcome Trust Centre for Human Genetics, during March 8-9, 2010. This kick-off meeting was important to clarify the main goals during the early phase of the study, to standardize the information to be collected in the case-control samples and to establish timelines and goals in the study. During the meeting, we had two guest speakers, who had expertise in statistical analyses (Prof. Esteban Parra) and admixture mapping (Prof. Paul MacKeigue). These experts shared their views on the challenges and opportunities of gene identification in the admixed populations from Latin America. This meeting was attended by all original Partners in the Consortium, with the exception of KBIOSCIENCE, who rarely engaged in the study.

The Second Consortium meeting took place in Cancun, during February 13-14, 2011.
Business discussions around samples and genotyping platforms predominated. A major decision of this meeting was the pronouncement, unanimously taken by the Consortium, to ask Argentina to leave the study due to lack of progress. Partners also gave selected background scientific presentations and guest speakers presented their work on disease gene identification in Latino populations (Drs. Josh. Galanter, Andres Moreno and Karla Sandoval).

The third Consortium Meeting was held at San Carlos Hospital in Madrid between March 19th and 21st 2012. The meeting was primarily dedicated to the sample collection reports from the Latin American Partners and to the design of the genotyping panel to be used in the study. Partner 1 (OU) presented the statistical genetic work carried out to design the second part of the Latino array. The Consortium agreed that this was the best option and agreed the principle of array design as described above. Guest speakers were also invited (Dr. Jose-Luis Gomez-Skarmeta) who presented his work on functional assessment of GWA SNPs. The Madrid CHIBCHA meeting was held back-to-back with that of the international colorectal cancer genetics consortium, COGENT, that is principally based on European groups. CHIBCHA members were formally incorporated as members of COGENT, and plans agreed to perform combined data analysis, for example of rare coding variants typed by both groups.

The Final meeting of the Consortium took place in Sacramento, USA, during June 22-25 2013. This meeting was mostly dedicated to discussion about the final results of the study, to the preparation of the final scientific report and to plans for the statistical analyses of the data. All Partners agreed to the final analysis plans and to the list of potential publications that will be derived from the study. During this meeting, and in collaboration with the University of California Genome Center (where the Project Manager currently has his research group), a well-attended mini-symposium on admixture mapping and cancer genetics took place and had guest speakers from CHIBCHA (Ian Tomlinson and Sergi Castellvi-Bel), the University of California San Francisco (Trevor Graham, Chris Gignoux and Elad Ziv) and the University of Southern California (Andrea Sottoriva).

In summary, the deliverables in WP1 and WP8 were achieved and the Project Manager successfully worked closely with the Co-ordinator office and all Partners to ensure that the work proposed in each of the WPs of the Consortium were carried out satisfactorily and to the highest standards.

2.1. Blood samples (Month 34) and 2.2. DNA samples (Month 36)

Four of the Partners (2, 3, 4, 9) have been recruiting samples for the study.

This part of the work plan was critical to the success of the subsequent molecular genetic studies. It includes case identification, control identification, specifics of sample collection, sample processing, collection of clinico-pathological data, and determination of family, medical and reproductive history. Cases were ascertained prospectively in order to minimize any hypothetical effects of differential survival on our analysis.

Colombia (Partner 2)

The patient recruitment was based on Hospital Federico Lleras Acosta, in Ibagué, Instituto Nacional de Cancerología in Bogotá, Hospital Pablo Tobón Uribe in Medellín, Hospital Hernando Moncaleano in Neiva, Registro Poblacional de Cancer in Pasto and others. To expand sample collection, we set up a network of 8 more collaborators in other additional Colombian cities, such as Bucaramanga.

We obtained the necessary ethical review board approval for patient ascertainment, recruitment, blood sampling and storage, DNA extraction and storage, collection of data from past medical history, family history, reproductive history, socio-economic status, alcohol and smoking, sending samples to collaborators and performing genetic analyses. We obtained also informed consent for the controls. Cases were contacted through pathologists, surgeons or oncologists. Controls were spouses or partners sampled through the proband, or obtained from true population cohorts. Brief questionnaires were answered by the patients or controls. These questionnaires have demographic and basic clinico-pathological data, as well as information about family history of CRC and other malignancies. Full, informed consent was obtained from each participant. In a few cases, the leader of the group obtained the ethical approval in the hospitals or other institutions to use anonymised samples. Copies of the Questionnaire forms and consent forms are under custody in the group and can be supplied on request. All data are held on a secure server. Access is username and password-protected and it is limited to prevent those who hold the key to anonymised samples from accessing molecular data. From cases suspected of having a Mendelian CRC syndrome, we requested tumour samples from histopathology archives.

Colombian blood samples July 2013

The Colombian cases and controls have supplied peripheral blood and DNA has been extracted from these locally and by Partner 1. These DNAs have been sent for genotyping to Partner 7.

Mexico Universidad Autónoma de Nuevo León Monterrey (Partner 3)

Mexican samples comprised incident colorectal cancer cases and population controls, with basic clinical data (age, sex) and pathological data (location and type of cancer). Blood samples were taken from all individuals, DNA was extracted and the samples were sent to Partner 1 and, after quantitation, from there to Partner 7 for genotyping.

Brazil (Partner 4)

Patients were recruited from the Sao Paulo Hospital AC Camargo Cancer Center, following approval in the IRB (local and national). The cases were recruited in the Departments of Oncology and Pelvic Surgery. All clinical and pathological data were reviewed from the electronic medical record. Controls were recruited in the Department of Community Medicine who perform a programme for early cancer detection in the communities of the city of Sao Paulo. Demographic data were obtained as for the Mexican samples in interviews at the time of recruitment. 54 cases and 50 controls were excluded because they were of Asian origin. Thus, 1049 cases 1000 controls remained. The clinical and demographic data are presented below.

Age at diagnosis of CRC cases (left) and controls (right)

Distribution of cases regarding tumour location and staging.

Variable N %
Tumor location
Right colon 238 23.9
Left colon 341 34.3
Rectum 398 40.0
Not known 18 1.8

0 22 2.2
I 144 14.5
II 252 25.3
III 274 27.5
IV 229 23.0
X 74 7.4

IPO-Porto Portuguese Oncology Institute, Porto (Partner 9)

Samples were 495 CRC and 5 colorectal adenoma patients (235 male, mean age at diagnosis 51; SD±11.81). About four-fifths of the patients were recruited according to the Bethesda or Amsterdam criteria for Lynch syndrome and had a negative genetic testing. The remaining patients were recruited with the inclusion criteria of the CHIBCHA project: diagnosis of either CRC at <75 years old, or of an ‘advanced’ CRA (villous histology, or >1cm diameter or severe dysplasia at <60 years old). The 500 controls (269 male, mean age 47; SD±8.7) were recruited for the CHIBCHA project from blood donors of the Portuguese Oncology Institute of Porto.

Resumé of sample collection activities

Overall, the recruiting Partners in the study generally reached sample collection expectations, albeit with some delays caused by several factors, critical amongst which was the failure of planned collections from two sites, Argentina and Ecuador, that did not occur for reasons never clarified by those ex-Partners. Attempts to substitute increased sample numbers from other sources were largely successful, but over a longer timescale than originally planned. The quality of DNA derived from blood samples has been very good, allowing the successful sequencing of known genes and SNPs in WP3 and the genotyping of SNP arrays in WP4 (see below).

The Colombian sample collection always proved more difficult than expected. The poor prognosis associated with CRC in the country has made it more difficult than predicted locally to recruit incident and prevalent cases. A large fraction of patients presented with late-stage disease and died during the patient identification process. The high mortality associated with CRC in the country is an important finding of the study and is, we believe, more likely to result from a failure of clinicians and patients to recognise, report and act on symptoms than the existence of intrinsically more aggressive disease. The poor prognosis of CRC in Colombia has, moreover, highlighted the need for early detection, for example through the establishment of CRC screening programs in the country. Partner 2 is currently in discussion with local policy makers to establish a pilot evaluation of bowel screening program in the region. In addition, the Colombian case and control collection has been hit by two recent factors: the relocation of the Project Manager, a native Colombian, to the USA; and the contamination of a large batch of DNA samples delivered by the Colombian Partner to Partner 1, about 1,000 blood samples had to be re-extracted by the Colombian group (Partner 2) and were delivered directly to the genotyping Partner (USC) in March 2013. This was more than 6 months later than the date proposed in the latest Technical Annex report (September 2012). Colombia remain about 800 samples short of their proposed total of 3,000, although this has entirely been compensated for by the better than planned recruitment from Mexico. In addition, Partner 2 has collected a panel of about 50 tumour-normal pair DNAs from Colombia, and an exceptional resource of nearly 700 Native American blood samples.

3.1 Known CRC genes in Latin American patients (Month 35)


Familial cases with histories suspicious of a Mendelian colorectal cancer syndrome were recruited in Brazil, Colombia and Mexico for this part of the CHIBCHA study. In these efforts, familial cases were grouped according to clinical and familial history suggestive of familial adenomatous polyposis type 1 caused by germline mutations in the APC gene (FAP 1, locus at 5q22.2) and type 2 caused by mutations in the MUTYH gene (FAP 2, locus at 1p34.1); and cases suggestive of hereditary nonpolyposis co-lorectal cancer (HNPCC) or Lynch syndrome, caused by germline mutations in the in mismatch repair (MMR) genes MSH2 (locus at 2p21), MLH1 (locus at 3p22.2); PMS2 (locus at 7p22.1) MSH6 (locus at 2p16.3) TGFBR2 (locus at 3p24.1) and MLH3 (locus at 14q24.3). Lynch syndrome cases were ascertained through screening using the Amsterdam and Bethesda criteria for HNPCC. As a result, the following unrelated cases of suspected inherited CRC were collected in each participating country:
Brazil: FAP, 23; Lynch syndrome, 109.
Colombia: FAP, 6; Lynch syndrome, 36; other polyposis, 3.
Mexico: FAP 12; Lynch syndrome, 96

Mutation screening

Molecular screening is still in progress in most of the participating countries. So far, these analyses were finished by the Brazilian participant and they published an article describing the mutational spectrum of mutations in the APC and MUTYH genes in FAP cases (Torrezan GT, et al). Of the 23 Brazilian FAP probands tested, 20 had mutations in the APC gene and 6 in the MUTYH gene.

Five novel mutations were found in the APC gene. Most of these mutations were distinct of those reported in Caucasian populations, suggesting no founder effects in the admixed Brazilian population, and common variants observed in Caucasians were absent or rarely found. Notoriously, a significant number of APC mutation-positive families were not consistent for predicted genotype-phenotype correlations. The Brazilian mutations in APC comprised:
c.856_859dupCATG (p.Glu287Alafs*2). Novel mutation
c.447dupC (p.Lys150Glnfs*18). Novel mutation
c.4097dupC (p.Gln1367Serfs*8). Novel mutation
c.5365G > C (p.Val1789Leu). Novel mutation
Exon 1–3 duplication. Novel mutation
c.3927-3931delAAAGA (p.Glu1309Aspfs*4)
c.3927-3931delAAAGA (p.Glu1309Aspfs*4)
del 5q21.3-q22.3 (chr5:107916475–113079330 Hg19)
c.904C > T (p.Arg302*)
c.4348C > T (p.Arg1450*)
c.3880-3881delCA (p.Gln1294Glyfs*6)
c.847C > T (p.Arg283*)
c.4099C > T (p.Gln1367*)
c.3050-3053delATGA (p.Asn1017Metfs*4)
c.4393-4394delAG (p.Ser1465Trpfs*3)

The Brazilian mutations in MUTYH comprised:
c.[536A > G]; [1147delC] (p.[Tyr179Cys]; [Ala385Profs*23])
c.[536A > G]; [1227-1228dup] (p.[Tyr179Cys]; [Glu410Glyfs*43])
c.[389-1G > C]; [536A > G] (p.[Val130GlufsX98;p.(spl?)]; [Tyr179Cys])
c.[1187G > C];[=] (p.[Gln396Asp];[=])
c.[348 + 33_*64 + 146del4285insTA]; [348 + 33_*64 + 146del4285insTA]
c.[721C > T]; [721C > T] (p.[Arg241Trp]; [Arg241Trp])

Probably pathogenic mutations in the Lynch syndrome genes were found in the Brazilian samples as follows:
MLH1: 15 cases
MSH2: 18 cases
MSH6: 1 case
PMS2: 2 cases

The Partners in Colombia, Oxford and Barcelona recently identified and published a founder MLH1 mismatch repair mutation in Lynch Syndrome patients from Antioquia in Colombia. A manuscript from this study (Alonso-Espinaco et al, 2011) was published in Genetics in Medicine in 2011 and the reference was provided in the first periodic report.

The Colombian Partner also carried out a clinico-pathological characterization of a subset of the Colombian samples. This study (Florez et al, 2012) was recently accepted for publication in the Colombian Journal of Gastroenterology.

In Mexico, molecular screening for mutations is in progress, but results are not yet available. Partner 3 has performed a comparison of the presence of the MLH1 and MSH2 proteins in tumour samples by immunohistochemistry (IHC) and by detecting the BRAF V600E mutation in 57 patients under 50 years old and 48 patients above years old. As expected, there was a significant association between defective MMR protein expression and young age. No instances of the BRAF V600E were found. These results were published (Luévano-González A, et al. Analysis of DNA mismatch repair proteins expression and BRAF V600E mutation in a subset of early- and late-onset colorectal carcinoma patients in Mexico. Arch Med Res. 2011; 42:457-462).

3.2. Assessing the known European colorectal cancer SNPs in Latin American cases and controls (Month 37)

Transfer of medium-throughput SNP genotyping technologies were effective from Wellcome-Trust (U.K.) Universidad de Santiago de Compostela (Spain), and currently, from University of California at Davis (U.S.) mainly to the laboratories of the Colombian and Mexican participants, and in lesser degree, to the participants in Brazil and Portugal, who implemented most of these technologies before the starting of the CHIBCHA project. This was performed mainly through the in-situ training of graduate students from the participating countries, to discussions in the annual CHIBCHA meetings, and to remote assessment/counselling.

All three Latin American Partners have also carried out work on the examination of already-established common, low-penetrance variants in the region. The Brazilian Partner has genotyped 5 SNPs in ~1,400 samples and is planning to visit one of the European Partners to receive training in the analysis of the data. The Colombian and Mexican Partners have also carried out evaluation of a number of the known SNPs (10 in Colombia and 6 in Mexico) in a subset of their samples. The proposed work on common SNPs in Colombia and Mexico has not, however, been extensively pursued, because all 20 known SNPs have been included in the genome-wide arrays that are currently being used to genotype ~4,000 samples (2,000 from Colombia and 2,000 from Mexico, see WP3, below). The analysis of these data is in progress and we plan to produce a manuscript describing these associations, and potentially fine mapping the European disease-causing SNPs, by the end of 2013.

The Portuguese Partner has additionally genotyped his samples for the following known CRC SNPs
• rs4444235 and rs1957636 at 14q22.2; rs961253 and rs4813802 at 20p12.3 (BMP2/BMP4) – finished (Carcinogenesis 34: 314-318)
• rs4779584 and P35A at 15q13.3 (GREM1) - finished
• rs11987193 at 8p12 (EPICOLON) - finished
• rs10791240 at 11q25 (COGENT) - finished
• rs12080929 at 1p33 (EPICOLON) - finished
• Exome arrays (~220K coding SNPs in 200 cases+200 controls) - analysis ongoing

4.1 5.1. GWAS and AMS genome-wide phases (Month 40)

Partner 7 genotyped a total of 45 96-well plates of samples, totalling 4,275 individuals from Mexico (2,041) and Colombia (2,234) in approximate 1:1 case:control ratios. The samples were typed for two panels comprising 1,200,749 SNPs throughout the whole genome. About 600,000 SNPs were derived from the Affymetrix Latin American panel and the remainder were custom content or proxy SNPs corresponding to SNP panels previously genotyped in European populations, specifically:
(ii) ~40,000 coding SNPs already validated as polymorphic in a large European population study;
(iii) additional SNPs from 1000genomes that have allele frequencies >5% in Latin America and <~1% in Europeans;
(iv) SNPs for cancer-related intermediate phenotypes such as height, body mass index, waist to hip ration, age at menopause and age at menarche;
(v) potentially pathogenic protein-coding variants of unknown significance from the mismatch repair genes, MUTYH and APC;
(vi) SNPs emerging from cancer patient/genome sequencing in Europeans.

For the Mexican samples, there were five collecting locations in Mexico, 3 in Monterrey, Nuevo Leon, 1 in Mexico DF and 1 in Torreón. A total of 2,722 individuals were sampled, 1,845 in Mexico DF, 51 in Torreón and 826 in Monterrey. For the genotyped samples, quality control filters were applied to the raw data and 4,070 individuals (>95%) passed QC, as did >99% of all SNPs.

SNP minor allele frequency distribution is shown below.

The ancestral proportions of the samples were estimated using STRUCTURE (version 2.3.4). The analysis was performed using AIMs from the LACE panel. Colours indicate: Red African, Green European, Blue Native-American. The first half of the admixed individuals corresponds to unaffecteds and the second half to affected individuals; note an apparent association of CRC with European ancestry proportion that is confirmed in the table below.

A triangle plot from STRUCTURE also shows varying degrees of Native American-European admixture in Mexico, with very little evidence of African ancestry. Cluster 1 corresponds to African, Cluster 2 to European and All others to Native-American. Yellow dots correspond to unaffected and pink dots correspond to affected individuals.

Principal component analysis using EIGENSTRAT detected evidence of population sub-structure in the Mexicans, almost certainly based on ancestry, but also reflected in the centre of collection.

An initial genome-wide association analysis has been performed on the Mexican samples. At this stage there has been no correction for ancestry (whether as estimated proportion or using principal components). The Manhattan plot below shows the results of an allelic test of association (-log10pallelic on y axis, and chromosome/position on x axis). Threshold Lines correspond to p-values 10-5 and 10-7.

The initial analysis of the Colombian samples has been performed separately from that of the Mexican samples. Most Colombian samples were collected from Antioquia, Bogota and Tolima, with 12% from other centres. 1,025 Colombian cases and 1,086 controls remained in the analysis after QC.

Estimated ancestry proportions were as shown. Colours indicate: Red African, Green European, Blue Native-American. The European and African proportions appear higher than in Mexico.

This is borne out by the triangle plot and the ancestry proportions.

Again, the affecteds have slightly more European ancestry.

Principal component analysis detected some evidence of population sub-structure. The causes are under investigation, but they probably primarily reflect ancestry, as expected.

A preliminary genome-wide association in Colombia (as for Mexico) above is shown below.

Full, ancestry-corrected association analyses will be performed under a variety of genetic models for both Mexico and Colombia in the next month. We shall assess systematic over-dispersion of the test statistic using QQ plots and the GC statistic. In order to reduce GC to an acceptable level (<1.10) It may be necessary to remove a small number of individuals from the analysis based on the principal component analysis. We shall then perform a meta-analysis of the two data sets, assess the origin of risk alleles (generally European or American), perform a comparison within our existing European GWAS data and select SNPs for validation typing in the Brazilian samples.

4.2. 5.2. GWAS and AMS validation phase (Month 42)

We plan to genotype approximately 400 SNPs on 3,000-4,000 samples, principally comprising the Brazilian cases and controls, supplemented by additional cases and controls from Colombia and Mexico that have not been typed on SNP arrays. We plan to genotype about 500 SNPs in each of these samples, comprising 300 best GWAS signals, 150 SNPs from the most promising regions of differential ancestry from the AMS, 50-100 AIMs, and a small number of candidate SNPs that have recently been found to show evidence of association with CRC in Europe, United States or Asia. Any GWAS signals that replicate will also be typed in the Portuguese cases and controls.

The consumables for this genotyping, which will be performed by Partner 7, have already been purchased. We anticipate that the choice of SNPs for this validation phase will be made in September 2013.

We have additionally instigated collaborations to perform a combined analysis of genetic predisposition to colorectal and endometrial cancers. the rationale for this is the fact that these cancers may share a common aetiology, through mismatch repair gene mutations in Lynch syndrome, through associations with intermediate phenotypes such as obesity, and through shared somatic mutation pathways (PTEN, PIK3CA, CTNNB1, KRAS, TP53). A preliminary European-based analysis has been performed, and a combined analysis with the Latin American data will be undertaken once these are fully available.

4.3 5.3. GWAS and AMS assessment (Month 42)

Once validation phase genotyping has been done, we shall perform the final GWAS and AMS analyses. We shall declare significant any GWAS association at P<5x10-8 and any AMS signal at LOD>3.0. In both cases, fine mapping in silico and combined analysis of the two types of data (see 6.3) will follow the discovery of any signal. We anticipate that this stage of the project will be completed in March 2014

4.4. Cancer sequencing (Month 40)

Approximately 25 cancers from Colombian cases will be sequenced for a panel of known cancer driver and candidate genes using the Ion Torrent (Ampliseq) cancer panel. The mutation spectrum will be compared against our existing data from 129 European colorectal cancers.

6.1. Databases (Month 8)

Sample and genotype databases have been set up and are in everyday use by Partners 1 and 10.

6.2. AIM panel testing (Month 36)

Galanter et al (2012), carried out an extensive validation of AIMs in 953 Native and admixed individuals from 18 Latin American populations from 9 countries and have produced the most comprehensive AIM panel to date for Hispanic Populations. The New AIM Panel was fully delivered by Partner 7 (USC), in collaboration with colleagues in the US-based Latin American Cancer Epidemiology (LACE) Consortium. We have used the panel to examine Colombian population genetics in three native groups (Pijao, Paez and Embera) and admixed populations from Tolima using mtDNA, Y chromosome markers and AIMs. Modern Colombians possess a complex genetic structure as a consequence of the admixture between Amerinds, Europeans and Africans in different proportions across regions of the country. We developed a genetic study aimed to characterize Amerindian, European and African ancestry in a Colombian Andean population (Tolima), which was compared with another one located at the Caribbean coast region (Córdoba).

Ethnicity Town
Total Samples
Samples used DNA extracted No. typed using AIMs panel
Naza-Paez Planadas-Tolima (Gaitania) 178 96 96 93
Embera-Katio Montelíbano and Pindo 34 17 17 17
Pijao Ortega-Tolima (Guatavita) 94 57 23 23
Pijao Coyaima-Tolima (Lomas de Hilarco) 65 25 25 25
Zenu Montelíbano-Cordoba (Porvenir La Fe) 116 82 20 20
Zenu Sahagun-Cordoba (km35) 139 42 30 24
Embera-Catio Rio Verde-Cordoba (Puerto Libertador) 26 5 5 4
Pijao Yaguara-Chaparral-Tolima 56 40 In progress 0
Embera-Chami Mariquita-Tolima 41 7 7 7

Self-identified indigenous people affiliated to 4 Native American ethnic groups from Tolima and Córdoba (n=248) were examined, together with 11 admixed samples. Non-indigenous or individuals without any ethnic affiliation (assumed as admixed) came from three municipalities of Tolima (n= 207) and one of Cordoba (n=31). Blood was sampled and DNA extracted by Partner 1. The founder Native American (A, B, C and D), African (L) and European (J) mtDNA haplogroups were typed by polymerase chain reaction (PCR/RFLP). Continental paternal lineages determined by the non-recombining part of the human Y chromosome (NRY) polymorphisms were identified through the 16 NRY-SNPs in total. We first typed haplogroup Q (M242: rs8179021) and R1 (M173: rs2032624) in all male plates, followed by E (M96: rs9306841), J (M304: rs13447352), I (M170: rs2032597), and G (M201: rs2032636) haplogroups in samples showing ancestral mutations for Q or R1. Samples having shown derived mutation at M242 and M173 were typed for their respective subhaplogroups: Q1a3* (M346), Q1a3a* (M3: rs3894), Q1a3a1 (M19: rs3910), Q1a3a2 (M194: rs2032677), R1b (M343: rs9786184), R1b1 (P25: rs150173), and R1b1b2 (M269: rs9786153). Some individuals negative for haplogroups above described, were tested for L (M22: rs3913), NO (M231: rs9341278) and for KR clade (M9: rs3900). Mutation M45 for P clade was typed to corroborate KR one in some samples only. A total of 100 of the most informative autosomal SNPs were selected from a previously developed panel of 400 markers generated for the LACE consortium to distinguish between the three parental populations from admixed individual in Latin American.

Maternal ancestral origin of each individual was specified based on the geographic distribution of the A, B, C, D, L and J mDNA haplogroups, as previously reported. For the Y chromosome, the paternal lineages were assigned according to the frequency and geographic origin of the examined NRY haplogroups, following the published literature. All frequencies and diversity index or between populations paired Fst were calculated using ARLEQUIN software, and derived Fst tables were used for multidimensional scaling analysis developed in GENALEX. Autosomal SNP genotypes were analysed using the Bayesian methods incorporated in STRUCTURE, implementing an admixture model aimed to identify ancestral populations. The analysis was performed without any prior population assignment The K (cluster node) value was set from 2 to 5, and was performed at least five times. Similarities between Native American, European or African ancestries were evaluated using multiple reference panels of ancestral populations (from Partner 7 or public databases).

Most individuals did not identify themselves with any specific native group. In the region of Tolima, Nasa and Pijao indigenous individuals were predominantly of Native American origin 91% (AIMs), 96% (mDNA) and 71% (NRY). Tolima admixed populations were paternally and maternally respectively 70% of European origin and 93 % Amerindian, and for the autosomal markers, 49% European and 45% Amerindian, with African ancestry <5%. Conversely, whilst the Embera indigenous population from the Caribbean coast had features similar to Tolima´s Amerinds, Zenu indigenous were ancestrally similar to admixed individual from the same region (European: AIMs 42%, NRY 67%; Amerind: AIMs 35%, NRY 10%, mtDNA 60%; African: AIMs: 23%, NRY: 24%, mtDNA: 13%). The results suggest that the whole sample is genetically structured, such that the last two sub-populations have a triparental origin, while Tolima´s admixed individuals are biparental, related not only the known higher African influence on the Caribbean coast, but also to specific local socio-cultural differences.

The inferred population structure, based on 1189 individuals from this and other studies, and genotype data from the 100 AIMs is as shown (yellow= European, purple=African, blue=Native American).

6.3. New GWA methods (Month 36)

We are currently testing the effects of explicitly incorporating admixture into GWA designs and comparing this with use of principal components. We are also assessing the feasibility of using a joint association-admixture statistic. This Deliverable will be worked on once genotyping is complete. It is anticipated in early 2014.

6.4. Allele-specific CNV method (Month 42)

We have developed this method for application to the germ line using as a model aneusomic cancer cell lines at which SNP allele dosage is unbalanced. The figure below shows the Kaspar method of genotyping for SNP rs7177. Blue and red dots represent homozygous samples. Green dots represent heterozygotes. For most heterozygotes, copy number is 2 and the points lie close to the blue line that represents equal allelic dosage (i.e. no copy number change). However, for samples 1206 and KTCL140, allelic dosage is unbalanced (2:1 or 1:2), indicating copy number gain.

6.5. Full statistical analysis (Month 48)

This Deliverable awaits completion of the GWAS and AMS genotyping in the Brazilian samples, followed by the basic GWAS and AMS results. We anticipate several “spin-off” studies over the months extending at least to the end of 2014. To take just one example, the Colombian samples have considerable data on phenotypes other than CRC (for example, anthropometric traits, other cancers). These samples may provide valuable insights into the genetics of these other traits, either as stand-alone projects or as parts of larger collaborative groups.

7.1. Basic functional data (Month 48)

Functional annotation of GWAS and AMS “hits” will be performed once 6.5 is complete. It is anticipated in mid-2014.

Future prospects

The Partners have all recently joined an EU COST action co-ordinated by Partner 8. This will ensure continuing collaboration, not least while the final genotyping of the Brazilian samples is performed, the data analysis is completed and various manuscripts written.

The planned timetable is as follows
Analysis of GWAS and AMS data from genome-wide SNP arrays August 2013
Selection of GWAS or AMS SNPs for validation in Brazilian samples September 2013
Genotyping of Brazilian and other samples November 2013
Final GWAS and AMS analysis March 2014
Functional annotation April 2014
Manuscripts May 2014

It must be emphasised that all the samples and the great majority of the data planned to be accrued under CHIBCHA have already been obtained.


Galanter JM, Fernandez-Lopez JC, Gignoux CR, Barnholtz-Sloan J, Fernandez-Rozadilla C, Via M, Hidalgo-Miranda A, Contreras AV, Figueroa LU, Raska P, Jimenez-Sanchez G, Zolezzi IS, Torres M, Ponte CR, Ruiz Y, Salas A, Nguyen E, Eng C, Borjas L, Zabala W, Barreto G, González FR, Ibarra A, Taboada P, Porras L, Moreno F, Bigham A, Gutierrez G, Brutsaert T, León-Velarde F, Moore LG, Vargas E, Cruz M, Escobedo J, Rodriguez-Santana J, Rodriguez-Cintrón W, Chapela R, Ford JG, Bustamante C, Seminara D, Shriver M, Ziv E, Burchard EG, Haile R, Parra E, Carracedo A; LACE Consortium. Development of a panel of genome-wide ancestry informative markers to study admixture throughout the Americas. PLoS Genet. 2012 Mar;8(3):e1002554. Epub 2012 Mar 8

Luévano-González A, Guzmán AQ, Ancer Rodríguez J, Ortiz López R, Rojas Martínez A, González Guerrero JF, Flores Gutiérrez JP. Analysis of DNA mismatch repair proteins expression and BRAF V600E mutation in a subset of early- and late-onset colorectal carcinoma patients in Mexico. Arch Med Res. 2011 Aug;42(6):457-62. Epub 2011 Sep 22.

Carvajal-Carmona LG, Zauber AG, Jones AM, Howarth K, Wang J, Cheng T; APC Trial Collaborators; APPROVe Trial Collaborators; CORGI Study Collaborators; Colon Cancer Family Registry Collaborators; CGEMS Collaborators, Riddell R, Lanas A, Morton D, Bertagnolli MM, Tomlinson I. Much of the genetic risk of colorectal cancer is likely to be mediated through susceptibility to adenomas. Gastroenterology. 2013 Jan;144(1):53-5

Jones AM, Beggs AD, Carvajal-Carmona L, Farrington S, Tenesa A, Walker M, Howarth K, Ballereau S, Hodgson SV, Zauber A, Bertagnolli M, Midgley R, Campbell H, Kerr D, Dunlop MG, Tomlinson I. TERC polymorphisms are associated both with susceptibility to colorectal cancer and with longer telomeres. Gut. 2012 Feb;61(2):248-54.

Tomlinson I, Carvajal-Carmona LG, Dobbins SE, Tenesa A, Jones AM, Howarth K, Palles C, Broderick P, Jaeger EE, Farrington S, Lewis A, Prendergast JG, Pittman AM, Theodoratou E, Olver B, Walker M, Penegar S, Barclay E, Whiffin N, Martin L, Ballereau S, Lloyd A, Gorman M, Lubbe S; COGENT Consortium; CORGI Collaborators; EPICOLON Consortium, Howie B, Marchini J, Ruiz-Ponte C, Fernandez-Rozadilla C, Castells A, Carracedo A, Castellvi-Bel S, Duggan D, Conti D, Cazier JB, Campbell H, Sieber O, Lipton L, Gibbs P, Martin NG, Montgomery GW, Young J, Baird PN, Gallinger S, Newcomb P, Hopper J, Jenkins MA, Aaltonen LA, Kerr DJ, Cheadle J, Pharoah P, Casey G, Houlston RS, Dunlop MG. Multiple common susceptibility variants near BMP pathway loci GREM1, BMP4, and BMP2 explain part of the missing heritability of colorectal cancer. PLoS Genet. 2011 Jun;7(6):e1002105. doi: 10.1371/journal.pgen.1002105. Epub 2011 Jun 2.

Houlston RS; members of COGENT (including Echeverry MM, Carvajal-Carmona, L, Tomlinson, I). COGENT (COlorectal cancer GENeTics) revisited. Mutagenesis. 2012 Mar;27(2):143-51. Review. PubMed PMID: 22294761; PubMed Central
PMCID: PMC3269000.

Alonso-Espinaco V, Giráldez MD, Trujillo C, van der Klift H, Muñoz J, Balaguer F, Ocaña T, Madrigal I, Jones AM, Echeverry MM, Velez A, Tomlinson I, Milà M, Wijnen J, Carvajal-Carmona L, Castells A, Castellví-Bel S. Novel MLH1 duplication identified in Colombian families with Lynch syndrome. Genet Med. 2011 Feb;13(2):155-60. PubMed PMID: 21233718.

Tomlinson I, Dunlop M, Campbell H, Zanke B, Gallinger S, Hudson T, Koessler T, Pharoah PD, Niittymäki I, Tuupanen S, Aaltonen LA, Hemminki K, Lindblom A, Försti A, Sieber O, Lipton L, van Wezel T, Morreau H, Wijnen JT, Devilee P, Matsuda K, Nakamura Y, Castellví-Bel S, Ruiz-Ponte C, Castells A, Carracedo A, Ho JW, Sham P, Hofstra RM, Vodicka P, Brenner H, Hampe J, Schafmayer C, Tepel J, Schreiber S, Völzke H, Lerch MM, Schmidt CA, Buch S, Moreno V, Villanueva CM, Peterlongo P, Radice P, Echeverry MM, Velez A, Carvajal-Carmona L, Scott R, Penegar S, Broderick P, Tenesa A, Houlston RS. COGENT (COlorectal cancer GENeTics): an international consortium to study the role of polymorphic variation on the risk of colorectal cancer. Br J Cancer. 2010 Jan 19;102(2):447-54. Epub 2009 Nov 17.

Torrezan GT, da Silva FC, Santos EM, Krepischi AC, Achatz MI, Aguiar S Jr, Rossi BM, Carraro DM. Mutational spectrum of the APC and MUTYH genes and genotype-phenotype correlations in Brazilian FAP, AFAP, and MAP patients. Orphanet J Rare Dis. 2013; 8:54.

Deliverables and milestones tables


The deliverables due in this reporting period, as indicated in Annex I to the Grant Agreement have to be uploaded by the responsible participants (as indicated in Annex I), and then approved and submitted by the Coordinator. Deliverables are of a nature other than periodic or final reports (ex: "prototypes", "demonstrators" or "others"). If the deliverables are not well explained in the periodic and/or final reports, then, a short descriptive report should be submitted, so that the Commission has a record of their existence. If a deliverable has been cancelled or regrouped with another one, please indicate this in the column "Comments". If a new deliverable is proposed, please indicate this in the column "Comments". This table is cumulative, that is, it should always show all deliverables from the beginning of the project.

Table 1. Deliverables

Delno. Deliverable name Version WP no. Lead beneficiary Nature Dissemination
level Delivery month (proj) Actual/Forecast delivery date Status Contract-ual? Comments
3.1 Known CRC genes in LA 2 3 3 O CO 34 31/12/2013 Not submitted No
3.2 Known CRC SNPs in LA 2 3 3 O CO 34 31/8/2013 Not submitted No
6.3 New GWA methods 2 6 1 O CO 36 30/11/2012 Not submitted No
6.2 AIM panel testing 2 5 7 O CO 36 30/6/2013 Submitted No
5.1 Native Amern genotyping 2 5 7 O CO 36 30/6/2013 Submitted No
2.2 DNA samples 2 2 2 O CO 36 31/3/2013 Submitted No
1.3 EU reports 2 1 1 R RE 36 31/7/2013 Submitted No
4.1 Genome-wide phase GWAS 2 4 1 O CO 40 31/8/2013 Not submitted No
5.1 Genome-wide phase AMS 2 5 7 O CO 40 31/8/2013 Not submitted No
4.4 Cancer sequencing 2 4 1 O CO 40 30/11/2013 Not submitted No
4.2 Validation phase GWAS 2 4 1 O CO 42 28/2/2014 Not submitted No
5.2 Validation phase AMS 2 5 7 O CO 42 28/2/2014 Not submitted No
4.3 Assessment of GWAS 2 4 1 O CO 42 31/3/2014 Not submitted No
5.3 Assessment of AMS 2 5 7 O CO 42 31/3/2014 Not submitted No
6.4 Allele-specific CNV method 2 6 1 O CO 42 31/3/2013 Submitted No
8.1 Reporting of Results 2 8 1 O CO 42 31/5/2014 on Not submitted No
8.2 Raising awareness 2 8 1 O CO 42 31/5/2014 on Not submitted No
8.3 Training LA researchers 2 8 1 O CO 42 30/6/2013 Submitted No
8.4 IP identification 2 8 1 O CO 42 31/5/2013 Not submitted No
8.5 CRC predisposition panel 2 8 1 O CO 42 30/4/2013 Not submitted No
7.1 Basic functional data 2 7 1 O CO 48 30/4/2013 Not submitted No
6.5 Full statistical analysis 2 6 1 O CO 48 End 2014 Not submitted No
1.3 EU report 2 1 1 R RE 48 31/7/2013 Submitted No


Please complete this table if milestones are specified in Annex I to the Grant Agreement. Milestones will be assessed against the specific criteria and performance indicators as defined in Annex I.

This table is cumulative, which means that it should always show all milestones from the beginning of the project.

List and schedule of milestones

Milestone number Milestone name Work package Lead beneficiary Expected date Comments

1.1 Each 6-monthly report of progress 1 1 6/12/18/24/30/36 Assessed on the basis of the 1-2 monthly reports
1.2 Successful quality control assessment 1 1 3/6/9/12/15/18/21/24/27/30/33/36/39/42 Assessed on the basis of the 1-2 monthly reports
2.1 Sample collection and processing following three-monthly target figures in each centre 2 2 3/6/9/12/15/18/21/24/27/30/33
2.2 Choice of genotyping platform 2 2 18
3.1 Successful running of mutation screening and SNP typing in local centres 3 3 34
3.2 Assessment of whether SNP typing has aided in refining locations of CRC genes identified with tagSNPs 3 3 35
4.1 Assessment of genotyping progress 4 1 24/27/30/33/36/39/42
4.2 Data analysis after each Phase, and SNP selection for subsequent Phase(s) 4 1 30/33/36/39
4.3 If required, contacting collaborators for “LA-only” SNPs 4 1 40
4.4 Planning for additional funding to take forward certain associations 5 1 39
5.1 Assessment of genotyping progress 5 7 24/27/30/33/36/39/42
5.2 Data analysis after each Phase, and region/SNP selection for subsequent Phase(s) 5 7 30/33/36/39
5.3 Usefulness of contacting collaborators for prime candidate SNPs 5 7 40
5.4 Planning for additional funding to take forward certain associations 5 1 39
6.1 Genotyping data for AIM map obtained and useable 6 11 24
6.2 Genotyping data for GWAS and AMS present in quantity and quality sufficient for analysis 6 11 30/33/36/39
7.1 Decision on which genes and proteins to assess on the basis of in silico data 7 1 41
8.1 Assessment of progress in dissemination and training 8 1 18/42

Potential Impact:

List of potential impacts

1. A new Latin American ancestry informative marker (AIM) panel has been developed by Partner 7 as part of a collaboration with groups in the USA. This panel has already been used and validated as part of CHIBCHA with great success.

2. Partner 2 (Colombia) has worked extensively with Native American tribes in the region. She has talked extensively to be tribes’ representatives about the project, in terms of the ancestry of the region and the growing importance of cancer, especially bowel cancer, as a growing health problem. She has made videos of these activities that are available from her on request.

4. Tests for the known colorectal cancer predisposition genes from European populations are now fully developed in the Latin American populations – these findings will allow the introduction of genetic testing for high-penetrance mutations into areas of previously low service provision and/or uptake, principally Colombia.

5. The known common polymorphisms associated with bowel cancer risk will be assessed in the Latin American populations. This will provide a direct assessment of the portability of these polymorphisms across populations and is likely to assist in identifying the functionally important variation close to each polymorphism. These activities may contribute to the development of new anticancer strategies, for example by targeted chemoprevention.

6. We have developed a new panel of polymorphisms, including some rare variants, that is tailored to use in Latin American populations. This panel builds on the commercially-available Affymetrix Axiom Latino array, using the latest data from public genotyping projects to ensure better coverage of the genome. In addition, content will be enriched for markers that are uncommon in European populations, but more common in Latin Americans. Finally, the array includes tens of thousands of rarer genetic variants (especially from protein-coding regions) that have been discovered by large sequencing projects in individual with Latin American ancestry.

7. The full GWAS analysis will be empowered to identify common colorectal cancer risk alleles that are hard to find in European populations. The significant (or highly promising) findings will also be examined in European and Asian populations, as well as in other cancers such as endometrial cancer, that share some genetic aetiology with bowel cancer.

8. The admixture mapping screen will deliver regions of the genome derived from Native American or European ancestors that are differentially represented in cases (and controls) compared with the overall ancestry proportions for each individual. This method is an important alterative strategy to cancer gene identification for alleles that cannot be detected by methods such as GWAS.

9. Ultimately, by completing more of the genetic jigsaw of bowel cancer risk, we will bring closer the time at which genetic risk can be used – ideally alongside demographics and environment – in a model of risk at the level of the general population. We have begun to address this issue already in the UK. Dr Ceres Fernandez has provided added value to CHIBCHA by obtaining a Marie Curie Fellowship to study whether common polymorphisms can be used in a cost-effective way to stratify screening for bowel cancer. Should this appear feasible in the UK, it can be rolled out to other countries, including the rest of Europe and Latin America.

10. A method for allele-specific copy number assessment using SNP typing on the ABI Taqman platform has been developed, based on fitting regression lines through data points generated by relative allele intensity data. This method has been used and published in a different context (Schödel et al. Nat Genet. 2012; 44: 420-5)

11. Somatic mutation next-generation sequencing data of a custom gene panel (Ion Torrent cancer panel) will be obtained from about 25 Latin American colorectal cancers. Comaprison will be made with our own and published data from European cancers. This work is likely to identify any mutations that are over-represented in Latin American cases, potentially suggesting selection of chemotherapy or prediction of outcome for individuals.

12. Consortium meetings with guest scientific speakers have been held in the UK, Mexico and Spain and USA, the latest being in June 2013 (all Partners). In one case (Madrid), the meeting was combined with that of the European colon cancer genetics consortium (COGENT), with considerable success and cross-fertilisation. Guest speakers were present at all meetings and provided excellent educational talks to general benefit. We have levered additional EU funding as a result of this collaborative work in the form of a COST grant led by Partner 8.

13. Secondment of researchers from Colombia and Uruguay to European Partners for training and knowledge transfer has occurred. There have been three visits of periods up to 3 months of Latin American students to the UK (2 from Colombia, 1 from Uruguay). Three Mexican students have also visited Santiago de Compostela for training (Partner 7). In addition, there has been continual transfer of technical know-how, especially to Colombia and data analysis know-how to Uruguay.

14. There has been education of local clinicians (especially in Colombia) in the genetics of colorectal cancer, the potential benefits of prevention through targeted screening and the symptoms of the disease. This has been achieved largely through recruitment of collaborators to the project from several of the major cities in Colombia. The profile of genetic testing for bowel cancer has also been raised in Mexico.

15. The aims and activities within CHIBCHA have been disseminated to the local scientific and popular media, especially in Latin America. Some examples are provided by:

16. It is important to note that much of the dissemination of the results of the project has not been possible yet owing to the overrun caused by delays to sample collection. Critically, all samples and the great majority of the molecular data have been obtained from genome-wide genotyping of the Colombian and Mexican samples. The molecular data awaited are from the validation phase of the GWAS and AMS. Once we have our findings, we shall disseminate them not only in the global scientific journals and on our website, but also in scientific journals and the relevant University websites in Latin America. We shall make every effort to engage the local media and populations, as has already been exemplified by Partner 2.

List of Websites:

Contact: Ian Tomlinson, Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, UK
Tel +44 (0)1865 287500
Fax +44 (0)1865 287502