Despite the new paradigm of hypothesis-generating science in the field of biological research, translating diverse types of large-scale experimental data into testable hypotheses is still quite challenging. Because C. elegans has highly effective gene perturbation methods including RNA interference and CRISPR-Cas9 knockout system, bioinformatics tools for prioritizing candidate gene-to-phenotype associations would facilitate genetics research of complex traits in animals.
Previously, we published network-assisted prediction servers based on genome-scale co-functional networks of C. elegans genes, WormNet v1 (Nature Genetics 40:181) and v2 (Genome Research 20:1143). WormNet v3 is a new prediction server with major updates to its base gene network as well as network analysis methods. The new base gene network showed substantially improved predictions of RNAi phenotypes collected from WormBase220(See the following plots).
Precision of network links were assessed by the percentage of gene pairs that share same annotations of RNAi phenotype over the entire range of the coding genome coverage. WormNet v3 clearly shows superior precision to that of the previous WormNet. The improved precision is attributable to superiority in precision of new network links for WormNet v3 (WormNet v3 specific links).