Final Report Summary - GENE2FARM (Next generation European system for cattle improvement and management)
Executive Summary:
The FP7 European research project Gene2Farm project finished in December 2015. For 48 months, 8 research partners addressed the needs of the cattle industry, in particular of 11 SMEs and end users, for an accessible, adaptable and reliable management system in their breeding programmes. As part of the project, a comprehensive and multi-level programme of work was undertaken that ranged from the development of statistical theory development and generation of new genomic data and knowledge, through the construction of practical and directly implementable tools. The project structure was defined by five scientific work-packages (WP1-WP5), one application WP (WP6), one training, transfer and dissemination WP (WP7), and a management WP (WP8).
The following briefly describes the objectives achieved by the project:
• Creating and testing new simulation procedures and novel statistical models and tools for within-, across- and multi-breed high-density genome selection.
• Addressing genetic diversity management and the optimization of breeding schemes, using new genomic information. Designing specific software for fast application of the designed procedures.
• Generating a large amount of genomic data: Gene2farm produced ~200 whole genome sequences, and more than 6000 genotypes of 14 dairy, beef, and dual-purpose European cattle populations.
• Designing a procedure to design specific SNP panels that capture the genetic diversity within and across breeds characterize the genomic variation and haplotypic structure of the breeds, and lead to an accurate parentage assessment.
• Assessing and producing a list of recommendations for the application at a large scale of current, new and future phenotyping protocols.
• Estimating genetic relatedness between populations using genomic (both sequencing and genotyping data) and pedigree information.
• Providing bioinformatics tools to facilitate SNP panel inter-changeability, which resulted in the publication of the SNPchimp open-access web interface (and linked software), which collects and links information from SNP array producers and public databases.
• Providing bioinformatics tools to analyze large datasets to: quality control, general data management, imputation and other analyses (Zanardi software), control inbreeding using genomic and phenotypic data (Optimate software) and to perform genomic evaluations (Gbcpp).
• Dissemination and know-how transfer activities implemented following four different strategies: 1) Project website and central information of the activities of the project; 2) Training courses for SME partners and their industry partners (from basic to advanced); 3) Training for scientists and SME collaborators from outside the project; 4) Establishing a SME Club of Interest was established to follow closely the progresses on the topics covered during the project.
• Validating all project outputs in real-case scenarios.
The collaboration and interaction among partners of the consortium led to the achievement of results far beyond the initial objectives of the project. It also contributed to obtain a more efficient use of resources in commercial breeding programs, by generating new knowledge and validating each of the results obtained.
Project Context and Objectives:
The main objective for the Gene2Farm project was to address the needs of the cattle industry, in particular of SMEs and end users, for an accessible, adaptable and reliable management system in their breeding programmes. As part of the project, a comprehensive and multi-level programme of work was undertaken that ranged from the development of statistical theory development and generation of new genomic data and knowledge, through the construction of practical and directly implementable tools.
The project structure was defined by five scientific work-packages (WP1-WP5), one application WP (WP6), one training, transfer and dissemination WP (WP7), and one management WP (WP8).
The overall aim of WP1 was to create novel statistical models and tools for high-density genome selection, along with developing and testing new software that would allow implementation of the tools and models developed. This WP focused entirely on simulated data, as it was considered not possible to generate a sufficiently high volume of genotyping and sequence data that was needed to underpin the work within the time and budget constraints of the project. Its specific objectives were to explore: i) the needs and targets of the SME partners, in order to efficiently address the project resources; ii) new models for genomic evaluation using whole-genome sequencing data; iii) the use of genomic information to facilitate the management of genetic diversity, considering that multiple breeds with multiple production systems were included in the project; iv) models and scenarios to evaluate the impact of single-, across- and multi-breed estimation of genomic breeding values; v) the impact of the introduction of new statistical models into breeding schemes, accounting for multiple plausible parameters, and; vi) a fine-scale model to evaluate genomic effects at chromosome level.
WP2 was focused on generating all the planned new genomic information for the project, studying the characteristics of the breeds involved and creating a set of tools to gain fast and user-friendly access to historical and new data. The overall approach adopted was: first to identify a set of key individuals of each cattle population involved in the project; then, to generate and analyze genotype and whole genome sequencing of the selected individuals. The basic outcomes of this WP were designed to underpin the work planned in all other WPs. The main focus of the analyses was understanding general sequence variation (generating general statistics for each of the breeds considered), SNP panel inter-changeability (generating full information on the SNP chips used during the project, to facilitate imputation), and identification of ancestral haplotypes, and methodologies to optimize marker panel designs.
WP3 focused on needs and practices regarding current and novel trait phenotypes. One of the main objectives of this WP was to produce a set of recommendations for the application of results in large-scale routine recording, including standardization methods for trait measurements. Firstly, the project’s SME were surveyed to update the state of the art on practices and needs. Secondly, on-farm recording, abattoir monitoring and novel phenotype collection approaches for both dairy and beef cattle were studied, and discussions held with other European/international organizations, with the aim of finding possible methods to solve some of the current difficulties of standardizing such data, and to release a list of recommendations.
Gene2farm also produced new knowledge on population structure research. Full genomic assessment of the breeds involved in the project was performed in WP4, with a special focus on signals of selection and differentiation between populations. Effective population size and structure (and variations) were fully assessed. The focus was on relationships and genetic distances among individuals and populations, which affect the success of across population genomic evaluations. Selection footprints, and haplotype conservation and differentiation were also assessed both on simulated and real data. Parameters influencing SNP panel number and distribution of markers for highly accurate imputation were also studied.
WP5 had a central role in this project, as the work focused on translating all methods and tools developed in the previous WPs into protocols and practical user-friendly tools to address partner’s SME needs. In fact this WP managed all the software and bioinformatics outputs able to exchange, manage, integrate, and analyze genotypic and phenotypic data. This WP not only tested the software on real data produced by the project, but also created the documentation of guidelines and procedures in order to provide SME with full information on each of the tools created.
The main objective of WP6 was to apply all the knowledge produced in the project to full-scale real-case scenarios. In fact, the outcomes of WP1-WP5 were validated and applied directly by many of the SME partners to their data on dairy, beef and dual-purpose populations of variable dimension.
Training, knowledge transfer and dissemination activities for the project were done as part of WP7. The main objectives included: creating and maintaining the project website; establishing an education programme for industry participants; training SME partners to use of the tools developed in WP1-WP5 and validated in WP6; disseminating the activities of the project to the general public thorough the organization of winter schools, and establishing an “Industry Club of Interest” to ensure durability of a support network.
Finally, the work in WP8 focused on managing the project to ensure that the planned activity were carried out efficiently, objectives were met following SME specifics, collaboration and communication was facilitated, reports were delivered on time and that there was effective coordination with the EC officials.
Project Results:
The Gene2farm consortium achieved all the planned objectives and deliverables for the project. The consortium has been very active and collaborative to perform the research, and then transfer scientific outcomes into practical tools readily applicable by SME partners.
WP1:
The activities of WP1 were the core of the project from the theoretical point of view. Although based on simulated data, results from this research was the starting point for many of the outcomes that were delivered to SME in the form of user-friendly software and tools. The first activity performed was to assess the SME specific needs and targets through a survey-based questionnaire. The results of the questionnaire showed that there was a wide range amongst the consortium SME partners in terms of their current capacity to incorporate genomic information in their breeding programs (D1.1); consequently, the aspirations of the SMEs for the project differed widely. The most common aspirations by the SME partners were to be able to better manage genetic and genomic diversity within and between breeds and to develop single, across- and multi-breed genomic evaluations. Following SME indications, a Bayesian variable selection algorithm, BayesP, was used to increase the accuracy of selection in comparison with GBLUP, the method currently used by SME for their genomic evaluation (D1.2). Subsequently, these methods were tested on a low- and high-dimensional set of markers (mimicking SNP-array and WGS data) on a multi-breed simulation, including also admixture. The results obtained (Objective n.1) clearly indicated the need for identifying dynamics of admixture and common ancestral haplotypes and highlighted the need for an in-depth knowledge of the genetic diversity of the populations studied. In fact, algorithms were developed within this WP to manage genetic diversity (D1.3 Objective n.2) using genomic information (which was confirmed being more informative than genealogical/pedigree information only). In addition, this WP developed algorithms to effectively combine genetic progress in economically important traits with inbreeding/biodiversity considerations for optimal management of genetic resources in small breeds within the context of a genomic breeding programme (D1.4 Objective n.3).
WP2:
The Gene2farm project identified and produced WGS and SNP array-based genomic information for a large number of individuals (D2.1 D2.2 Objective n.4).
Sequencing strategy followed an average 7x sequencing coverage (e.g. 40% higher than anticipated in the DoW). This allowed 7 RTD partners and 1 SME partner to enter the 1000 bulls genome project (Run5) and gain access to considerable amount of sequence data for research. In total, Gene2farm sequenced: 49 Brown Swiss bulls, 50 Simmental/Fleckvieh (31 Fleckvieh, 16 Simmental and 3 Pezzata Rossa Italiana bulls), 20 Guernsey bulls and 27 Norwegian Red bulls (21 of which were provided by GENO). Also 48 Spanish breed individuals were sequenced: 7 of Parda de la Montana (4 single and 3 pooled samples), 5 Pirenaica (1 single and 4 pooled samples), 7 Toro de Lidia bulls (1 single and 6 pooled samples), 5 poolled Retinta bulls, 7 Avilena bulls (2 single and 5 pooled), 10 Albera bulls (1 single and 9 pooled) and 6 Rubia Gallega bulls (1 single and 5 pooled samples) and 1 (single) Bruna dels pireneus. Also the genotyping strategy was modified from the DoW, in agreement with SME partners. The samples genotyped were (with various densities): 2348 samples from the Fleckvieh gene pool; 1460 samples from the Brown Swiss gene pool; 159 individuals from the Simmental gene pool; 453 individuals within the Norwegian Red gene pool; and 1637 individuals within the Guernsey gene pool.
All sequencing data was analyzed to identify SNP and CNV polymorphisms using state of the art bioinformatics pipelines (D2.3). Furthermore, basic statistics were calculated (e.g. allele frequency spectrum, haplotype detection, linkage disequilibrium estimation) resulting in a large within- and across-population variability observed. In addition, the study on the potential impact of the use of sequence data on between breed predictions indicated that there is not much support that sequence data may have a big impact on evaluation accuracy, unless some prior information on causative mutations is considered (D2.5). This confirmed the findings from simulated data analyses in WP1.
Sequence data was also used to identify an innovative method to design a between- and across-breed panel using Gene2farm breeds. The procedure included markers that capture the genetic diversity within and across breeds characterize the genomic variation and haplotypic structure of the breeds, and lead to an accurate parentage assessment (D2.6 Objective n.5). In addition, the selection of markers had the constraint of keeping markers present in commercial panels to ensure high imputation accuracy and SNP marker integration. In fact, SNP panel inter-changeability was an important issue for most researchers at the beginning of this project. Collaborating directly with commercial panel producers and public databases, Gene2farm solved this problem for six major animal genetic communities by creating SNPchimp, an online open-access and user-friendly tool to store all this information in a standardized format and content (D2.4).
WP3:
Gene2farm assessed the current phenotyping practices among SME partners in order to identify current and future problems and opportunities for the different types of cattle breeds (D3.1). As for current practices, dairy production and certain functional traits are the most important in dairy breed selection indexes, with few traditional functional traits being widely recorded and evaluated for. Similarly, beef traits and conformation traits are the most important in beef breeds. However, direct health traits are only recorded in very few cases and practically never in beef breeds. The needs in terms of future priorities from SME are quite heterogeneous. To allow flexibility and wide coverage, this research was further extended to consider automated or semi-automated systems for on-farm trait recording (D3.2). Such type of measurements could fill the gap of current direct health trait collection. However, a major difficulty is the standardization of historical and current data across breeds to a common scale (D3.3). Although no “gold standard” is currently available, Gene2farm developed a multiple-trait approach to tackle the problem, which works well for most of the (historical) traits, even with the inclusion of automated system data. Of course, this system has some constraints and prerequisites (e.g. requires good correlations amongst automated systems and good pedigree connections between individuals recorded). At the same time, standardization of novel traits present new challenges that are yet to be solved, mainly due to the difficulty in centralizing the information at farm level and IP issues in the methodologies used to measure/collect such data (D3.4). Similar problems are faced when dealing with abattoir data, where main difficulties are: the centralization of information (e.g. collection, transfer and maintenance of a database), measurement variability (including classifier training) and variability in the type and number of phenotypes collected (D3.5). Gene2farm developed a set of recommendations for the application of the above results, indicating also a set of priorities/needs, opportunities and challenges (D3.6 Objective n.6). Most important recommendations are: i) to increase emphasis on novel traits, especially those deemed important by the industry (e.g. direct health traits, utilization efficiency, animal welfare and behavior and environmental impact traits); ii) scientific research should focus on the genetic and phenotypic characterization of these novel traits.; iii) on-farm recording of novel traits collected with automated/semi-automated system should be promoted; iv) collaborative initiatives aimed at merging datasets should be encouraged, as this is essential to obtain progress in historical or novel data management; v) software and methods, as those produced by Gene2farm, are essential to harmonize and standardize the currently available data; vi) a centralized collection of standardized information, including standardized procedures, data units and measurements will facilitate cattle management and improvement programmes within and across countries; vii) research on integrating novel traits and technologies into modern breeding programmes is expected and encouraged; viii) international collaboration under the coordination of the International Committee for Animal Recording and the International Bull Evaluation Service (Interbull) will ensure continuity and consistency in trait definition and phenotype usage across countries.
WP4:
All Gene2farm breed and population structures were explored using multiple techniques, with a wide range of objectives. The effective population size (Ne) was assessed using multiple methods. Runs of Homozygosity (ROH) based estimates gave quite variable results depending on the methods used. Including chromosome-specific recombination rates in Sved’s method gave similar estimates and trajectories to those from genomic and pedigree data analyses. LD and haplotype structures of cattle populations were assessed, providing useful information of haplotype persistency across generations in real data at variable marker density. Gene2Farm examined the measurement of relatedness, both in simulated and real data and assessed how commonly used measures of relationship might be affected by the availability of sequence data (D4.1 Objective n.7). Levels of relatedness between simulated populations appeared not to be significantly affected by the proportion of the genome under selection or differences between breeding objectives or combinations between the level of admixture and heritability of the selection index. However, changes in selection intensity may create a strong shift pattern in the relatedness between populations. Real data analysis was based on SNP array and whole genome sequence data, in general supporting the findings from simulated data analyses. In fact, the relatedness estimated with the whole genome sequence data showed a high correlation with the estimates obtained from the high density SNP chip for the whole population or only between the individuals within or between breed. The Gene2farm dataset underwent a full selection sweep analysis, which included: a) analysis between populations; b) frequency spectrum tests; c) haplotype length based tests; d) local genetic variation test (D.4.2). Results indicated that old selection or adaptation processes occurred prior to breed differentiation, detected by the site frequency spectrum or local genetic variation methods. The isolation of the populations generated the signals of selection identified by the differentiation methods and finally, the haplotype length method identified more recent selection events that are mostly specific to each population. In addition, Gene2farm created a method to identify ancestral haplotypes that was assessed on simulated and real data (D4.3). Results portrayed the different selection processes and histories in the various breeds/populations, evidencing differential evolutionary processes, identifying preserved regions possibly responsible for adaptation and resilience, assessing genetic relatedness between populations and potentially improving the genomic evaluation and selection practices within and across breed/population. Finally, Gene2farm developed a strategy to optimize SNP panel size by using a prior knowledge of LD and LD profiles of the populations (D4.4). Since singletons were relevant to the size of the final panels, improving the procedure would allow gaining knowledge on the role of the singletons in the population of interests.
WP5:
The bioinformatics tools developed in Gene2farm have facilitated the central management and exchange of all key outcomes and data of the project (D5.1 Objective n.8).
In addition, this WP developed a series of user-friendly tools (SNPchiMp v.3 Pedda_ROW, Pedda_MATRIX, iConvert, SNPConvert,Optimate, GBCPP_pipeline, SNP2CARRIER) to store, convert, standardize, integrate, impute and analyse genomic data (D5.2 D5.4 D5.5 and Objectives n.8 n.9 and n.10).
It is important to underline the success of the SNPchiMp online toole, also as means of dissemination of the project. This tool, developed within Gene2farm, enabled implementation of such tools to other animal species. With nearly 100 users per week and more than 300k downloads and queries since its very first release, this user-friendly database-linked website has quickly become the central repository for SNP array information in the whole animal genomics community. A suite of computer programs has been developed to enable genotype-guided breeding management programs for the individual industry partners. An algorithm based on R scripts and standardised descriptors was developed to merge cow phenotypes collected on the farm manually or automatically into standard databases for use in management and large-scale genetic improvement programmes (D5.3). Finally, Zanardi suite was developed as an all-in-one tool to integrate data and project and third-party software in a single user-friendly solution (D5.2).
In order to ensure full know-how transfer of the tools and knowledge produced in the project, full documentation and protocol guidelines were produced (D5.6).
WP6:
The collaboration between partners in WP5 and WP6 allowed testing all software developed in WP5 on real-case scenarios in dairy, dual-purpose and numerically small breeds. The results obtained confirmed the excellence of the software suitability as well as the validity of outcomes from all other WPs (D6.1 and D6.2 Objective n.11).
WP7:
Training and knowledge transfer was a key aspect of the Gene2farm project. A website was produced early in the project and regularly updated in order to share key information with the public (D7.1). Important internal and public information was also advertised through the website, linking the outputs of the project to specific menus (e.g. publications, dissemination activities and slides). Two SME training courses were organized to share and demonstrate the software produced in WP5 and tested in WP6 (D7.2). In addition, an industry targeted basic education programme was developed that SME partners could use to deliver their own training workshops, the slides from which were also made publically available on the project website so as to enhance the output of the project (D7.3). Two successful Winter Schools were held in 2013 and 2015, resulting in project activities and results being disseminated to over 100 participants from 23 countries of 5 continents (D7.4). Finally, in order to ensure the legacy of the project in the future, an industry-led Club of Interest was created (D7.5).
Potential Impact:
A constant objective during the duration of Gene2farm has been to make its results readily accessible to the SME partners and, in the long term, to the entire animal genetics community. This was achieved by regular WP meetings (virtual or physical) that involved all partners (both RTDs and SME) in order to enhance coordination and updates, and improve the project’s usefulness to the whole community. All tools produced within the project were originally structured to be easily retrieved and fully open-access.
Thanks to the breadth and depth of scientific results obtained within Gene2farm and the bioinformatics tools provided to the community, the impacts of this project cover all areas of interest to SME: from phenotype production, standardization and integration at a national and international level, to the standardization and integration of genomic data from multiple sources, to the personalized- and easy-access to the large datasets produced.
The great success of the initiatives kick-started by this project is demonstrated by the collective Gene2farm output and the way it is being used as base for further development:
- The setting up of a reference population for the Guernsey cattle has allowed this breed to start a genomic selection process that is now running in collaboration with internationals partners and is fully sustainable.
- The large genomic (SNP array and sequence data) information produced has generated unprecedented knowledge of the population characteristics of the breeds involved in Gene2farm. Such information will also be the basis of future research and development of tools and knowledge.
- The legacy of the project is ensured by the high quality of the outputs produced and the interaction with external projects (e.g. 1000 bulls genome project), local initiatives (e.g. collaboration of WP3 results with INRA task force on phenotyping), and future projects (e.g. Zanardi software is currently being upgraded through other relevant national and international projects). In addition, the Club of Interest has created a network of SME that will maintain the awareness of the challenges faced and possible solutions in the future.
WP1:
A full survey on the needs of the SME involved in the project was produced at the beginning of the project. The outcome of such activity evidenced the diversity in needs for genomic information amongst the SME involved in the project. Therefore, the Gene2farm project partnership had to face different problems almost on a case-by-case basis. This prompted WP1 activities to address general problems from different angles, parameters and perspectives. The Whole Genome Sequencing (WGS) and multi-population simulation software, fully open-source and shared on public repositories, is now a resource for the whole animal genetics community. The algorithm developed to calculate genomic breeding values from high-density genotyping (HD SNP arrays and WGS) was tested on small- and large-scale scenarios, yielding results that were later on tested on real datasets. The algorithm that allows for single- and two-step approaches, and a full assessment of the dissection of SNP effects at chromosome level was shared with SME partners. In addition, algorithms for genetic diversity studies and the implementation of breeding schemes based on a multiple set of variables were developed and tested on simulated data. All algorithms developed were essential to the activities of the project. In fact, all codes were integrated into a full-size computer suite (WP5), tested for compatibility and shared with SME partners to be applied in official genomic evaluations (WP6).
WP2:
One of the most important tasks of the Gene2farm project was to produce new genomic data for a number of European breeds. This objective led to the genotyping of nearly 7,000 individuals at variable density and the production of whole genome sequencing data for around 200. The impact of such large datasets was multi-faceted. The setting up of a reference population for the Guernsey cattle has allowed this breed to start a genomic selection process that is now fully developed, sustainable and run in collaboration with internationals partners. Full assessment of the genetic variability of small Spanish breeds generated unprecedented information on the diversity of these beef breeds. The genotypes produced by this project increased significantly the reference population of the Italian Pezzata Rossa. Furthermore, results of Fleckvieh/Pezzata Rossa/Simmental and Brown Swiss analyses enabled to test full implementation of across-breed genomic evaluations on a high-density SNP array panel and WGS data, assessing imputation from low to high density and from high density to full sequence. In addition, this data was used to design low-density Gene2farm SNP panels specific to individual breeds, which could add value to the research and commercial activities of the SME partners involved.
The large amount of WGS data allowed Gene2farm partners enter the 1000 bull genome consortium, the international repository of cattle WGS data. The true impact of this collaboration is long lasting, as SME will benefit from interaction with a large number of international researchers also after the end of the Gene2farm project.
The multiple-species, open-access online tool SNPchimp has quickly become the reference for SNP array related data in the whole animal genomics community. On average, 100 different users from all over the world visit and download the data contained in the online database each week, for a total of more than 9,000 entries and 300K downloads since its first release. This tool was awarded the E!mpowered logo by Ensembl.
WP3:
The impact of the activities of WP3 can be summarized with the outcome of its last deliverable, which outlines the significant results of each tasks. The list of recommendations, priorities, needs, opportunities and challenges in terms of phenotyping activities were shared with ICAR. In addition, this WP shared still confidential results (after the agreement of SME and RTD partners) with the “Survey of European and International infrastructures and projects related to phenotyping of farm animals”, lead by JF Hocquette (INRA). All activities of this WP involved SME partners as leaders of tasks, which means that all activities were performed within the framework of ongoing routine activities, and in close collaboration with breeders.
WP4:
The activities of this WP were focused on the production of scientific knowledge. All SME participated to the discussions of methods and results, disseminating the outcomes to their breeders in local dissemination activities. A deeper knowledge of the genetic structure and relationships between European cattle populations was produced helping to define the future selection and conservation strategies. Further, a global map of signatures of selection on these populations was produced, allowing the identification of targets of natural or artificial selection that shaped the current genetic configuration of European Cattle. Identified selection signals were linked to candidate genes hypothetically associated with the traits related to both natural and artificial selection. Moreover, the identification of ancestral haplotypes opens new possibilities for the implementation of across-breed genomic prediction strategies. A new procedure to optimize SNP panels provides a useful tool for expanding genotyping to a wider spectrum of individuals (or populations).
WP5 and WP6:
These WPs translated the know-how generated in the previous WPs into practical tools. In fact, a series of bioinformatics tools were produced to collect, share, standardize and integrate large datasets (genotypes from SNP arrays, WGS data, phenotypes and pedigrees) following strict security protocols. These tools were not only handy for the realization of the project, but also allowed central and easy access to the data and will be kept active after the project’s end.
A series of accessory tools to the SNPchimp database were released to help standardizing formats and allele coding, and to update map information from multiple reference genome assemblies. A set of R tools was developed to merge phenotypic data locally to internationally. The Zanardi pipeline, open-source and multi-platform (Linux and Mac) allows interacting with multiple third party and other Gene2farm-developed software and algorithms. The suite does not only allow integrating data from multiple files and formats (705 and PLINK) but it also allows to inter-connect and analyze all the software developed in the project that span from advance genomic evaluation models to optimization of breeding schemes based on pedigree or genomic information. The OPTIMATE suite of tools is an efficient tool to control inbreeding using genomic and phenotypic data i.e. preserving genetic diversity, as supported everywhere by current EU agricultural politic. GBCPP software allows implementing a full genomic evaluation using innovative genomic models. The source-codes were first presented at internal training courses and then to the public within the last Gene2farm Winter School (both, activities of WP7).
All the activities performed in WP5 were assessed and fully tested on real-case scenarios and routine evaluations in dairy, dual-purpose and numerically small breeds, confirming the tools developed within Gene2farm have and will have a large impact on the activities of the SME participating to the project.
WP7:
This was the dissemination and know-how transfer WP of the project, which was implemented following four different strategies: 1) Project website and central information of the activities of the project. The website was constantly updated, with all the information of the relevant activities, dissemination and publication. In fact the website currently contains links to the 9 peer-reviewed journals and 21 presentations in courses, workshops, meetings and congresses; 2) Training courses for SME partners and their industry partners (from basic to advanced). Two internal training courses were organized to transfer the know-how of the project and several one-to-one interaction sessions were put in place during the project. Gene2farm also produced basic training courses for the SME partners; 3) Training for scientists and SME collaborators from outside the project, through the organization of two winter schools. Both winter schools had a great success, being fully booked months in advance by participants from all five continents. In order to ensure the legacy of the project after its end, a Club of Interest was established to follow closely the progresses on the topics covered during the project.
List of Websites:
The website address is the following: http://www.gene2farm.com/
Relevant contact details – Gene2farm Coordinator:
Dr. Ezequiel Luis Nicolazzi
Fondazione Parco Tecnologico Padano
Via Einstein – Loc. Cascina Codazza
29600, Lodi (LO), Italy
Tel: +39 0371 4662221
E-mail: ezequiel.nicolazzi@ptp.it
The FP7 European research project Gene2Farm project finished in December 2015. For 48 months, 8 research partners addressed the needs of the cattle industry, in particular of 11 SMEs and end users, for an accessible, adaptable and reliable management system in their breeding programmes. As part of the project, a comprehensive and multi-level programme of work was undertaken that ranged from the development of statistical theory development and generation of new genomic data and knowledge, through the construction of practical and directly implementable tools. The project structure was defined by five scientific work-packages (WP1-WP5), one application WP (WP6), one training, transfer and dissemination WP (WP7), and a management WP (WP8).
The following briefly describes the objectives achieved by the project:
• Creating and testing new simulation procedures and novel statistical models and tools for within-, across- and multi-breed high-density genome selection.
• Addressing genetic diversity management and the optimization of breeding schemes, using new genomic information. Designing specific software for fast application of the designed procedures.
• Generating a large amount of genomic data: Gene2farm produced ~200 whole genome sequences, and more than 6000 genotypes of 14 dairy, beef, and dual-purpose European cattle populations.
• Designing a procedure to design specific SNP panels that capture the genetic diversity within and across breeds characterize the genomic variation and haplotypic structure of the breeds, and lead to an accurate parentage assessment.
• Assessing and producing a list of recommendations for the application at a large scale of current, new and future phenotyping protocols.
• Estimating genetic relatedness between populations using genomic (both sequencing and genotyping data) and pedigree information.
• Providing bioinformatics tools to facilitate SNP panel inter-changeability, which resulted in the publication of the SNPchimp open-access web interface (and linked software), which collects and links information from SNP array producers and public databases.
• Providing bioinformatics tools to analyze large datasets to: quality control, general data management, imputation and other analyses (Zanardi software), control inbreeding using genomic and phenotypic data (Optimate software) and to perform genomic evaluations (Gbcpp).
• Dissemination and know-how transfer activities implemented following four different strategies: 1) Project website and central information of the activities of the project; 2) Training courses for SME partners and their industry partners (from basic to advanced); 3) Training for scientists and SME collaborators from outside the project; 4) Establishing a SME Club of Interest was established to follow closely the progresses on the topics covered during the project.
• Validating all project outputs in real-case scenarios.
The collaboration and interaction among partners of the consortium led to the achievement of results far beyond the initial objectives of the project. It also contributed to obtain a more efficient use of resources in commercial breeding programs, by generating new knowledge and validating each of the results obtained.
Project Context and Objectives:
The main objective for the Gene2Farm project was to address the needs of the cattle industry, in particular of SMEs and end users, for an accessible, adaptable and reliable management system in their breeding programmes. As part of the project, a comprehensive and multi-level programme of work was undertaken that ranged from the development of statistical theory development and generation of new genomic data and knowledge, through the construction of practical and directly implementable tools.
The project structure was defined by five scientific work-packages (WP1-WP5), one application WP (WP6), one training, transfer and dissemination WP (WP7), and one management WP (WP8).
The overall aim of WP1 was to create novel statistical models and tools for high-density genome selection, along with developing and testing new software that would allow implementation of the tools and models developed. This WP focused entirely on simulated data, as it was considered not possible to generate a sufficiently high volume of genotyping and sequence data that was needed to underpin the work within the time and budget constraints of the project. Its specific objectives were to explore: i) the needs and targets of the SME partners, in order to efficiently address the project resources; ii) new models for genomic evaluation using whole-genome sequencing data; iii) the use of genomic information to facilitate the management of genetic diversity, considering that multiple breeds with multiple production systems were included in the project; iv) models and scenarios to evaluate the impact of single-, across- and multi-breed estimation of genomic breeding values; v) the impact of the introduction of new statistical models into breeding schemes, accounting for multiple plausible parameters, and; vi) a fine-scale model to evaluate genomic effects at chromosome level.
WP2 was focused on generating all the planned new genomic information for the project, studying the characteristics of the breeds involved and creating a set of tools to gain fast and user-friendly access to historical and new data. The overall approach adopted was: first to identify a set of key individuals of each cattle population involved in the project; then, to generate and analyze genotype and whole genome sequencing of the selected individuals. The basic outcomes of this WP were designed to underpin the work planned in all other WPs. The main focus of the analyses was understanding general sequence variation (generating general statistics for each of the breeds considered), SNP panel inter-changeability (generating full information on the SNP chips used during the project, to facilitate imputation), and identification of ancestral haplotypes, and methodologies to optimize marker panel designs.
WP3 focused on needs and practices regarding current and novel trait phenotypes. One of the main objectives of this WP was to produce a set of recommendations for the application of results in large-scale routine recording, including standardization methods for trait measurements. Firstly, the project’s SME were surveyed to update the state of the art on practices and needs. Secondly, on-farm recording, abattoir monitoring and novel phenotype collection approaches for both dairy and beef cattle were studied, and discussions held with other European/international organizations, with the aim of finding possible methods to solve some of the current difficulties of standardizing such data, and to release a list of recommendations.
Gene2farm also produced new knowledge on population structure research. Full genomic assessment of the breeds involved in the project was performed in WP4, with a special focus on signals of selection and differentiation between populations. Effective population size and structure (and variations) were fully assessed. The focus was on relationships and genetic distances among individuals and populations, which affect the success of across population genomic evaluations. Selection footprints, and haplotype conservation and differentiation were also assessed both on simulated and real data. Parameters influencing SNP panel number and distribution of markers for highly accurate imputation were also studied.
WP5 had a central role in this project, as the work focused on translating all methods and tools developed in the previous WPs into protocols and practical user-friendly tools to address partner’s SME needs. In fact this WP managed all the software and bioinformatics outputs able to exchange, manage, integrate, and analyze genotypic and phenotypic data. This WP not only tested the software on real data produced by the project, but also created the documentation of guidelines and procedures in order to provide SME with full information on each of the tools created.
The main objective of WP6 was to apply all the knowledge produced in the project to full-scale real-case scenarios. In fact, the outcomes of WP1-WP5 were validated and applied directly by many of the SME partners to their data on dairy, beef and dual-purpose populations of variable dimension.
Training, knowledge transfer and dissemination activities for the project were done as part of WP7. The main objectives included: creating and maintaining the project website; establishing an education programme for industry participants; training SME partners to use of the tools developed in WP1-WP5 and validated in WP6; disseminating the activities of the project to the general public thorough the organization of winter schools, and establishing an “Industry Club of Interest” to ensure durability of a support network.
Finally, the work in WP8 focused on managing the project to ensure that the planned activity were carried out efficiently, objectives were met following SME specifics, collaboration and communication was facilitated, reports were delivered on time and that there was effective coordination with the EC officials.
Project Results:
The Gene2farm consortium achieved all the planned objectives and deliverables for the project. The consortium has been very active and collaborative to perform the research, and then transfer scientific outcomes into practical tools readily applicable by SME partners.
WP1:
The activities of WP1 were the core of the project from the theoretical point of view. Although based on simulated data, results from this research was the starting point for many of the outcomes that were delivered to SME in the form of user-friendly software and tools. The first activity performed was to assess the SME specific needs and targets through a survey-based questionnaire. The results of the questionnaire showed that there was a wide range amongst the consortium SME partners in terms of their current capacity to incorporate genomic information in their breeding programs (D1.1); consequently, the aspirations of the SMEs for the project differed widely. The most common aspirations by the SME partners were to be able to better manage genetic and genomic diversity within and between breeds and to develop single, across- and multi-breed genomic evaluations. Following SME indications, a Bayesian variable selection algorithm, BayesP, was used to increase the accuracy of selection in comparison with GBLUP, the method currently used by SME for their genomic evaluation (D1.2). Subsequently, these methods were tested on a low- and high-dimensional set of markers (mimicking SNP-array and WGS data) on a multi-breed simulation, including also admixture. The results obtained (Objective n.1) clearly indicated the need for identifying dynamics of admixture and common ancestral haplotypes and highlighted the need for an in-depth knowledge of the genetic diversity of the populations studied. In fact, algorithms were developed within this WP to manage genetic diversity (D1.3 Objective n.2) using genomic information (which was confirmed being more informative than genealogical/pedigree information only). In addition, this WP developed algorithms to effectively combine genetic progress in economically important traits with inbreeding/biodiversity considerations for optimal management of genetic resources in small breeds within the context of a genomic breeding programme (D1.4 Objective n.3).
WP2:
The Gene2farm project identified and produced WGS and SNP array-based genomic information for a large number of individuals (D2.1 D2.2 Objective n.4).
Sequencing strategy followed an average 7x sequencing coverage (e.g. 40% higher than anticipated in the DoW). This allowed 7 RTD partners and 1 SME partner to enter the 1000 bulls genome project (Run5) and gain access to considerable amount of sequence data for research. In total, Gene2farm sequenced: 49 Brown Swiss bulls, 50 Simmental/Fleckvieh (31 Fleckvieh, 16 Simmental and 3 Pezzata Rossa Italiana bulls), 20 Guernsey bulls and 27 Norwegian Red bulls (21 of which were provided by GENO). Also 48 Spanish breed individuals were sequenced: 7 of Parda de la Montana (4 single and 3 pooled samples), 5 Pirenaica (1 single and 4 pooled samples), 7 Toro de Lidia bulls (1 single and 6 pooled samples), 5 poolled Retinta bulls, 7 Avilena bulls (2 single and 5 pooled), 10 Albera bulls (1 single and 9 pooled) and 6 Rubia Gallega bulls (1 single and 5 pooled samples) and 1 (single) Bruna dels pireneus. Also the genotyping strategy was modified from the DoW, in agreement with SME partners. The samples genotyped were (with various densities): 2348 samples from the Fleckvieh gene pool; 1460 samples from the Brown Swiss gene pool; 159 individuals from the Simmental gene pool; 453 individuals within the Norwegian Red gene pool; and 1637 individuals within the Guernsey gene pool.
All sequencing data was analyzed to identify SNP and CNV polymorphisms using state of the art bioinformatics pipelines (D2.3). Furthermore, basic statistics were calculated (e.g. allele frequency spectrum, haplotype detection, linkage disequilibrium estimation) resulting in a large within- and across-population variability observed. In addition, the study on the potential impact of the use of sequence data on between breed predictions indicated that there is not much support that sequence data may have a big impact on evaluation accuracy, unless some prior information on causative mutations is considered (D2.5). This confirmed the findings from simulated data analyses in WP1.
Sequence data was also used to identify an innovative method to design a between- and across-breed panel using Gene2farm breeds. The procedure included markers that capture the genetic diversity within and across breeds characterize the genomic variation and haplotypic structure of the breeds, and lead to an accurate parentage assessment (D2.6 Objective n.5). In addition, the selection of markers had the constraint of keeping markers present in commercial panels to ensure high imputation accuracy and SNP marker integration. In fact, SNP panel inter-changeability was an important issue for most researchers at the beginning of this project. Collaborating directly with commercial panel producers and public databases, Gene2farm solved this problem for six major animal genetic communities by creating SNPchimp, an online open-access and user-friendly tool to store all this information in a standardized format and content (D2.4).
WP3:
Gene2farm assessed the current phenotyping practices among SME partners in order to identify current and future problems and opportunities for the different types of cattle breeds (D3.1). As for current practices, dairy production and certain functional traits are the most important in dairy breed selection indexes, with few traditional functional traits being widely recorded and evaluated for. Similarly, beef traits and conformation traits are the most important in beef breeds. However, direct health traits are only recorded in very few cases and practically never in beef breeds. The needs in terms of future priorities from SME are quite heterogeneous. To allow flexibility and wide coverage, this research was further extended to consider automated or semi-automated systems for on-farm trait recording (D3.2). Such type of measurements could fill the gap of current direct health trait collection. However, a major difficulty is the standardization of historical and current data across breeds to a common scale (D3.3). Although no “gold standard” is currently available, Gene2farm developed a multiple-trait approach to tackle the problem, which works well for most of the (historical) traits, even with the inclusion of automated system data. Of course, this system has some constraints and prerequisites (e.g. requires good correlations amongst automated systems and good pedigree connections between individuals recorded). At the same time, standardization of novel traits present new challenges that are yet to be solved, mainly due to the difficulty in centralizing the information at farm level and IP issues in the methodologies used to measure/collect such data (D3.4). Similar problems are faced when dealing with abattoir data, where main difficulties are: the centralization of information (e.g. collection, transfer and maintenance of a database), measurement variability (including classifier training) and variability in the type and number of phenotypes collected (D3.5). Gene2farm developed a set of recommendations for the application of the above results, indicating also a set of priorities/needs, opportunities and challenges (D3.6 Objective n.6). Most important recommendations are: i) to increase emphasis on novel traits, especially those deemed important by the industry (e.g. direct health traits, utilization efficiency, animal welfare and behavior and environmental impact traits); ii) scientific research should focus on the genetic and phenotypic characterization of these novel traits.; iii) on-farm recording of novel traits collected with automated/semi-automated system should be promoted; iv) collaborative initiatives aimed at merging datasets should be encouraged, as this is essential to obtain progress in historical or novel data management; v) software and methods, as those produced by Gene2farm, are essential to harmonize and standardize the currently available data; vi) a centralized collection of standardized information, including standardized procedures, data units and measurements will facilitate cattle management and improvement programmes within and across countries; vii) research on integrating novel traits and technologies into modern breeding programmes is expected and encouraged; viii) international collaboration under the coordination of the International Committee for Animal Recording and the International Bull Evaluation Service (Interbull) will ensure continuity and consistency in trait definition and phenotype usage across countries.
WP4:
All Gene2farm breed and population structures were explored using multiple techniques, with a wide range of objectives. The effective population size (Ne) was assessed using multiple methods. Runs of Homozygosity (ROH) based estimates gave quite variable results depending on the methods used. Including chromosome-specific recombination rates in Sved’s method gave similar estimates and trajectories to those from genomic and pedigree data analyses. LD and haplotype structures of cattle populations were assessed, providing useful information of haplotype persistency across generations in real data at variable marker density. Gene2Farm examined the measurement of relatedness, both in simulated and real data and assessed how commonly used measures of relationship might be affected by the availability of sequence data (D4.1 Objective n.7). Levels of relatedness between simulated populations appeared not to be significantly affected by the proportion of the genome under selection or differences between breeding objectives or combinations between the level of admixture and heritability of the selection index. However, changes in selection intensity may create a strong shift pattern in the relatedness between populations. Real data analysis was based on SNP array and whole genome sequence data, in general supporting the findings from simulated data analyses. In fact, the relatedness estimated with the whole genome sequence data showed a high correlation with the estimates obtained from the high density SNP chip for the whole population or only between the individuals within or between breed. The Gene2farm dataset underwent a full selection sweep analysis, which included: a) analysis between populations; b) frequency spectrum tests; c) haplotype length based tests; d) local genetic variation test (D.4.2). Results indicated that old selection or adaptation processes occurred prior to breed differentiation, detected by the site frequency spectrum or local genetic variation methods. The isolation of the populations generated the signals of selection identified by the differentiation methods and finally, the haplotype length method identified more recent selection events that are mostly specific to each population. In addition, Gene2farm created a method to identify ancestral haplotypes that was assessed on simulated and real data (D4.3). Results portrayed the different selection processes and histories in the various breeds/populations, evidencing differential evolutionary processes, identifying preserved regions possibly responsible for adaptation and resilience, assessing genetic relatedness between populations and potentially improving the genomic evaluation and selection practices within and across breed/population. Finally, Gene2farm developed a strategy to optimize SNP panel size by using a prior knowledge of LD and LD profiles of the populations (D4.4). Since singletons were relevant to the size of the final panels, improving the procedure would allow gaining knowledge on the role of the singletons in the population of interests.
WP5:
The bioinformatics tools developed in Gene2farm have facilitated the central management and exchange of all key outcomes and data of the project (D5.1 Objective n.8).
In addition, this WP developed a series of user-friendly tools (SNPchiMp v.3 Pedda_ROW, Pedda_MATRIX, iConvert, SNPConvert,Optimate, GBCPP_pipeline, SNP2CARRIER) to store, convert, standardize, integrate, impute and analyse genomic data (D5.2 D5.4 D5.5 and Objectives n.8 n.9 and n.10).
It is important to underline the success of the SNPchiMp online toole, also as means of dissemination of the project. This tool, developed within Gene2farm, enabled implementation of such tools to other animal species. With nearly 100 users per week and more than 300k downloads and queries since its very first release, this user-friendly database-linked website has quickly become the central repository for SNP array information in the whole animal genomics community. A suite of computer programs has been developed to enable genotype-guided breeding management programs for the individual industry partners. An algorithm based on R scripts and standardised descriptors was developed to merge cow phenotypes collected on the farm manually or automatically into standard databases for use in management and large-scale genetic improvement programmes (D5.3). Finally, Zanardi suite was developed as an all-in-one tool to integrate data and project and third-party software in a single user-friendly solution (D5.2).
In order to ensure full know-how transfer of the tools and knowledge produced in the project, full documentation and protocol guidelines were produced (D5.6).
WP6:
The collaboration between partners in WP5 and WP6 allowed testing all software developed in WP5 on real-case scenarios in dairy, dual-purpose and numerically small breeds. The results obtained confirmed the excellence of the software suitability as well as the validity of outcomes from all other WPs (D6.1 and D6.2 Objective n.11).
WP7:
Training and knowledge transfer was a key aspect of the Gene2farm project. A website was produced early in the project and regularly updated in order to share key information with the public (D7.1). Important internal and public information was also advertised through the website, linking the outputs of the project to specific menus (e.g. publications, dissemination activities and slides). Two SME training courses were organized to share and demonstrate the software produced in WP5 and tested in WP6 (D7.2). In addition, an industry targeted basic education programme was developed that SME partners could use to deliver their own training workshops, the slides from which were also made publically available on the project website so as to enhance the output of the project (D7.3). Two successful Winter Schools were held in 2013 and 2015, resulting in project activities and results being disseminated to over 100 participants from 23 countries of 5 continents (D7.4). Finally, in order to ensure the legacy of the project in the future, an industry-led Club of Interest was created (D7.5).
Potential Impact:
A constant objective during the duration of Gene2farm has been to make its results readily accessible to the SME partners and, in the long term, to the entire animal genetics community. This was achieved by regular WP meetings (virtual or physical) that involved all partners (both RTDs and SME) in order to enhance coordination and updates, and improve the project’s usefulness to the whole community. All tools produced within the project were originally structured to be easily retrieved and fully open-access.
Thanks to the breadth and depth of scientific results obtained within Gene2farm and the bioinformatics tools provided to the community, the impacts of this project cover all areas of interest to SME: from phenotype production, standardization and integration at a national and international level, to the standardization and integration of genomic data from multiple sources, to the personalized- and easy-access to the large datasets produced.
The great success of the initiatives kick-started by this project is demonstrated by the collective Gene2farm output and the way it is being used as base for further development:
- The setting up of a reference population for the Guernsey cattle has allowed this breed to start a genomic selection process that is now running in collaboration with internationals partners and is fully sustainable.
- The large genomic (SNP array and sequence data) information produced has generated unprecedented knowledge of the population characteristics of the breeds involved in Gene2farm. Such information will also be the basis of future research and development of tools and knowledge.
- The legacy of the project is ensured by the high quality of the outputs produced and the interaction with external projects (e.g. 1000 bulls genome project), local initiatives (e.g. collaboration of WP3 results with INRA task force on phenotyping), and future projects (e.g. Zanardi software is currently being upgraded through other relevant national and international projects). In addition, the Club of Interest has created a network of SME that will maintain the awareness of the challenges faced and possible solutions in the future.
WP1:
A full survey on the needs of the SME involved in the project was produced at the beginning of the project. The outcome of such activity evidenced the diversity in needs for genomic information amongst the SME involved in the project. Therefore, the Gene2farm project partnership had to face different problems almost on a case-by-case basis. This prompted WP1 activities to address general problems from different angles, parameters and perspectives. The Whole Genome Sequencing (WGS) and multi-population simulation software, fully open-source and shared on public repositories, is now a resource for the whole animal genetics community. The algorithm developed to calculate genomic breeding values from high-density genotyping (HD SNP arrays and WGS) was tested on small- and large-scale scenarios, yielding results that were later on tested on real datasets. The algorithm that allows for single- and two-step approaches, and a full assessment of the dissection of SNP effects at chromosome level was shared with SME partners. In addition, algorithms for genetic diversity studies and the implementation of breeding schemes based on a multiple set of variables were developed and tested on simulated data. All algorithms developed were essential to the activities of the project. In fact, all codes were integrated into a full-size computer suite (WP5), tested for compatibility and shared with SME partners to be applied in official genomic evaluations (WP6).
WP2:
One of the most important tasks of the Gene2farm project was to produce new genomic data for a number of European breeds. This objective led to the genotyping of nearly 7,000 individuals at variable density and the production of whole genome sequencing data for around 200. The impact of such large datasets was multi-faceted. The setting up of a reference population for the Guernsey cattle has allowed this breed to start a genomic selection process that is now fully developed, sustainable and run in collaboration with internationals partners. Full assessment of the genetic variability of small Spanish breeds generated unprecedented information on the diversity of these beef breeds. The genotypes produced by this project increased significantly the reference population of the Italian Pezzata Rossa. Furthermore, results of Fleckvieh/Pezzata Rossa/Simmental and Brown Swiss analyses enabled to test full implementation of across-breed genomic evaluations on a high-density SNP array panel and WGS data, assessing imputation from low to high density and from high density to full sequence. In addition, this data was used to design low-density Gene2farm SNP panels specific to individual breeds, which could add value to the research and commercial activities of the SME partners involved.
The large amount of WGS data allowed Gene2farm partners enter the 1000 bull genome consortium, the international repository of cattle WGS data. The true impact of this collaboration is long lasting, as SME will benefit from interaction with a large number of international researchers also after the end of the Gene2farm project.
The multiple-species, open-access online tool SNPchimp has quickly become the reference for SNP array related data in the whole animal genomics community. On average, 100 different users from all over the world visit and download the data contained in the online database each week, for a total of more than 9,000 entries and 300K downloads since its first release. This tool was awarded the E!mpowered logo by Ensembl.
WP3:
The impact of the activities of WP3 can be summarized with the outcome of its last deliverable, which outlines the significant results of each tasks. The list of recommendations, priorities, needs, opportunities and challenges in terms of phenotyping activities were shared with ICAR. In addition, this WP shared still confidential results (after the agreement of SME and RTD partners) with the “Survey of European and International infrastructures and projects related to phenotyping of farm animals”, lead by JF Hocquette (INRA). All activities of this WP involved SME partners as leaders of tasks, which means that all activities were performed within the framework of ongoing routine activities, and in close collaboration with breeders.
WP4:
The activities of this WP were focused on the production of scientific knowledge. All SME participated to the discussions of methods and results, disseminating the outcomes to their breeders in local dissemination activities. A deeper knowledge of the genetic structure and relationships between European cattle populations was produced helping to define the future selection and conservation strategies. Further, a global map of signatures of selection on these populations was produced, allowing the identification of targets of natural or artificial selection that shaped the current genetic configuration of European Cattle. Identified selection signals were linked to candidate genes hypothetically associated with the traits related to both natural and artificial selection. Moreover, the identification of ancestral haplotypes opens new possibilities for the implementation of across-breed genomic prediction strategies. A new procedure to optimize SNP panels provides a useful tool for expanding genotyping to a wider spectrum of individuals (or populations).
WP5 and WP6:
These WPs translated the know-how generated in the previous WPs into practical tools. In fact, a series of bioinformatics tools were produced to collect, share, standardize and integrate large datasets (genotypes from SNP arrays, WGS data, phenotypes and pedigrees) following strict security protocols. These tools were not only handy for the realization of the project, but also allowed central and easy access to the data and will be kept active after the project’s end.
A series of accessory tools to the SNPchimp database were released to help standardizing formats and allele coding, and to update map information from multiple reference genome assemblies. A set of R tools was developed to merge phenotypic data locally to internationally. The Zanardi pipeline, open-source and multi-platform (Linux and Mac) allows interacting with multiple third party and other Gene2farm-developed software and algorithms. The suite does not only allow integrating data from multiple files and formats (705 and PLINK) but it also allows to inter-connect and analyze all the software developed in the project that span from advance genomic evaluation models to optimization of breeding schemes based on pedigree or genomic information. The OPTIMATE suite of tools is an efficient tool to control inbreeding using genomic and phenotypic data i.e. preserving genetic diversity, as supported everywhere by current EU agricultural politic. GBCPP software allows implementing a full genomic evaluation using innovative genomic models. The source-codes were first presented at internal training courses and then to the public within the last Gene2farm Winter School (both, activities of WP7).
All the activities performed in WP5 were assessed and fully tested on real-case scenarios and routine evaluations in dairy, dual-purpose and numerically small breeds, confirming the tools developed within Gene2farm have and will have a large impact on the activities of the SME participating to the project.
WP7:
This was the dissemination and know-how transfer WP of the project, which was implemented following four different strategies: 1) Project website and central information of the activities of the project. The website was constantly updated, with all the information of the relevant activities, dissemination and publication. In fact the website currently contains links to the 9 peer-reviewed journals and 21 presentations in courses, workshops, meetings and congresses; 2) Training courses for SME partners and their industry partners (from basic to advanced). Two internal training courses were organized to transfer the know-how of the project and several one-to-one interaction sessions were put in place during the project. Gene2farm also produced basic training courses for the SME partners; 3) Training for scientists and SME collaborators from outside the project, through the organization of two winter schools. Both winter schools had a great success, being fully booked months in advance by participants from all five continents. In order to ensure the legacy of the project after its end, a Club of Interest was established to follow closely the progresses on the topics covered during the project.
List of Websites:
The website address is the following: http://www.gene2farm.com/
Relevant contact details – Gene2farm Coordinator:
Dr. Ezequiel Luis Nicolazzi
Fondazione Parco Tecnologico Padano
Via Einstein – Loc. Cascina Codazza
29600, Lodi (LO), Italy
Tel: +39 0371 4662221
E-mail: ezequiel.nicolazzi@ptp.it