Abstract

CAGI is a community experiment to objectively assess computational methods for predicting the phenotypic impacts of genomic variation, particularly in areas relevant to human disease. The aim of the experiment is to define the state-of-the-art of genome interpretation, both highlighting innovation and identifying those bottlenecks that prevent the field from advancing. In these experiments, participants are invited to make bona fide blinded predictions of disease-related phenotypes on the basis of genomic data.

The variation data used in challenges mirror that encountered in current and imminent clinical practice, with a focus on genomes, exomes, nonsynonymous SNPs, splice-affecting SNPs, and copy number variation, as well as additional data such as transcriptomes. Challenges probe the state of the art in identifying causative variants in rare Mendelian diseases and estimating the risk of individuals having complex diseases, such as Crohn’s. Cancer related variants form the basis of a substantial fraction of the challenges, relevant to the rapidly growing use of sequencing data for analysis and optimal treatment.