COSI - VarI

Advances in high-throughput sequencing and genotyping, as well as experiments aimed at the characterization of haplotype diversity and linkage disequilibrium are consistently generating vast amounts of genomic variation information in need of analysis. Genomic variation is particularly interesting in the context of its phenotypic manifestations, e.g. disease. For some monogenic diseases, such as cystic fibrosis and sickle-cell anemia, the association with the respective variants is well documented. In most other cases, the genomic causes of disease are still unknown. Genome-wide association studies (GWAS) provide some information about the variation-disease relationships. However, GWAS are both experimentally and computationally expensive and either poorly cover, or do not cover at all, the rare variants. The discrepancy between the significant availability of genomic data and the lack of its clinical usefulness spurs the development of algorithms for prediction of mutation impacts. Currently, a number of methods exist to prioritize variants for GWAS, annotate effects of SNVs, and/or define highly conserved genomic regions across different species and populations. Despite these efforts, a reliable method for predicting variome-to-phenotype associations (phenotype associations that consider the full set of variants in the genome) is not yet available. Arguably, part of the reason for the absence of such a holistic approach to variation analysis is that there is no clear "winner" among the methods available for mutation impact prediction. Moreover, it is not obvious for most researchers what are the differences between the available tools and in what context should one or the other be used.

In the near future, the study of the relationships between genetic variation and disease, in association with pharmaco-genomics, will be key factors for the development of personalized medicine. Many initiatives have been taken to create a forum for discussion of genomic variation analysis. In 2007 the ISMB Steering Committee introduced into the conference proceedings a session on Bioinformatics and Disease. The Critical Assessment of Genomic Interpretation experiment will take place for the fourth time this year (CAGI2015) to evaluate the efficacy of the available computational methods for predicting the phenotypic impacts of genomic variation. This year's VarI-SIG focus will mainly be on computational methods for variome interpretation. We've invited researchers interested in generation and annotation of large-scale variant repositories and the development of corresponding analysis tools. We specifically reached out to investigators working at the interface of computational tool development and method applications in medical context (particular focus in the "bridging the data gap" section of the main SIG).

Taking a step back from whole variome analysis, however, VarI-SIG will also hold an extended discussion session (morning of July 12th, 2015) within the Disease thematic area, one of the four thematic areas of the main ISMB/ECCB conference. The topic for this session will be something that interests us very much, both as scientists and as methods developers: what is the REAL difference between the computational methods meant to predict effects of genetic variation? If the performance of each individual method can no longer be attributed to data availability, what is it that drives the difference? Is it computational techniques? Better annotations? Better understanding of the interaction of the underlying forces, e.g. evolution, experimental active site annotations, 3D-structure? To find out, we invited researchers that have participated in the creation of the latest variant prediction/annotation methods - Graham Ritchie (University of Edinburgh), Martin Kircher (University of Washington, Seattle), Colin Campbell (University of Bristol), and Olivier Lichtarge (Baylor College of Medicine). The format of this session will include a brief introduction for each of the methods presented by their creators and an hour-long discussion between the discussion participants, VarI-SIG main-meeting keynote speakers, and, potentially, the audience.

We expect an audience for this session to consist of the researchers interested in all aspects of genomic variation research. Because of the variety that such study entails, we anticipate the discussion to take us into areas such as sequence analysis, protein structure & function, protein interactions & molecular networks, transcriptomics & gene regulation, disease models & epidemiology, population genomics & evolution and comparative genomics. We welcome participants that do not attend the main SIG meeting as well.

A SIG addressing the important issues in the interpretation of genetic variants (VarI-SIG) is necessary for organization of a research network facilitating the exchange of ideas and the establishment of new collaborations bringing together varying expertise. This year we will also aim to formalize the presence of a VarI-COSI [very cozy] community of scientists interested in variation analysis. We will use the first twenty minutes of the session, prior to the speaker presentations, to update all attendees on the current state of the community affairs and to ask for input in future organization of activities. The organization of the VarI-COSI community is a necessary point on the road towards understanding of the genomic information; a route that will require an unprecedented collaborative effort to manage the complexity of the analysis and evaluation of genetic variation.