Fig. 4

Performance comparison of VarElect to competing tools. a Phenolyzer: Thirty four queries were submitted to analysis in both VarElect and Phenolyzer (Methods). Each query is based on real results and includes a disease causing gene spiked into background genes, accompanied with real disease related symptoms that were recorded from the relevant patients (Table S1). The rank of the probe gene is color coded, bottom to top: top 10, top 11–20, top 21–100, below top 100, not found to be connected to the search terms. b Exomiser: Ten queries were submitted to both VarElect and Exomiser (Methods). The rank of the probe gene resulting from each query in Exomiser was recorded from 2 different score types; 1) Exomiser gene phenotype score that is solely derived from gene to phenotype relevance 2) Exomiser general score, combining variant severity score and gene phenotype score. Ranking results are shown with binning as in Fig. 4a. c Ingenuity: Queries as in Fig. 4b were submitted to Ingenuity Variant Analysis, and the appearance of the spiked gene in the resultant gene list was examined. For this comparison, Ingenuity 0-hop mode was considered comparable to VarElect direct mode, and Ingenuity 1-hop mode was considered comparable to VarElect indirect mode. d Phevor: Queries as in Fig. 4b were submitted to Phevor2 in combination with VAAST3 filtering (Methods). Ranking results are shown with binning as in Fig. 4a