Fig. 2: SpliceAI and Pangolin programs predict bona fide SDV in cattle.
From: Functional impact of splicing variants in the elaboration of complex traits in cattle
a A positive set of 24 SDV previously described in vivo and associated with phenotypes was compiled from PubMed and OMIA databases. b SpliceAI and Pangolin scores representing the probability to disrupt splicing were calculated for each variant classified by variant consequence. Variants with a score ≥0.2 (high recall threshold), ≥0.5 (recommended threshold), or ≥0.8 (high precision threshold) were predicted to be SDV as proposed by ref. 60. Variants reported to be associated with a monogenic disease are marked by an asterisk; others are related to complex traits. Note that Pangolin scores are represented in absolute values to allow comparison with SpliceAI scores. Syn synonymous, Mis missense, 3′ss 3′ splice site, 5′ss 5′ splice site, P Int proximal intronic, D Int distal intronic. c Proportion of variants predicted to be spliceogenic or non-spliceogenic classified by variant consequence or d phenotype. Graphs displaying stacked green and purple bars indicate SpliceAI and Pangolin predictions, respectively. The number of variants for each score category is shown on the Y-axis, with the highest categories shown in the darkest colours. Variants with a score equal to or higher than 0.2 are predicted to be SDV. Pangolin scores are represented in absolute values. Source data are provided as a Source Data file.
