Skip to main content
Fig. 2 | Genome Medicine

Fig. 2

From: Genome-wide prediction of pathogenic gain- and loss-of-function variants from ensemble learning of a diverse feature set

Fig. 2

Structure- and sequence-based protein feature analysis. a Enrichments and depletions for protein structural and functional features used by the LoGoFunc model. GOF (blue), LOF (orange), and neutral (green) log odds ratios are displayed for each feature. Significant enrichments and depletions are denoted by asterisks. Significance was calculated with Fisher’s exact test, Benjamini–Hochberg corrected [44] to allow for multiple comparisons. (Left) Features derived from protein sequences or protein interaction data. (Right) Features derived from AlphaFold2 [12] protein structures. b AlphaFold2 predicted structure of the Vasopressin V2 receptor protein. (Left) Residues colored by the number of HGMD [15] pathogenic variants occurring in the nine closest neighboring residues in space. (Right) Residues colored by the number of gnomAD [17] variants occurring in the nine closest neighboring residues in space

Back to article page

Genome Medicine

ISSN: 1756-994X

Contact us