Skip to main content
Fig. 1 | Genome Biology

Fig. 1

From: Introme accurately predicts the impact of coding and noncoding variants on gene splicing, with clinical applications

Fig. 1

Assessment of existing in silico splice prediction tools on functionally validated splice-altering variants. A The performance of each tool is shown using precision recall curves (PRCs) of functionally validated splice-altering and non-splice-altering variants, grouped by the class of the variant. PRCs compare the precision (proportion of calls that are relevant) and recall (proportion of total relevant variants that are called) across different score thresholds, with better performing tools appearing near the top right. Variants that impacted the polypyrimidine tract or branchpoint were classified as impacting the acceptor splice site (3′SS). 5′SS: donor splice site; ESE/ESS: exonic Splicing Enhancer/Silencer. B Comparison of variants identified by splicing tools. The overlap of known splice-altering variants that are also predicted to be splice-altering by each in silico tool (i.e., above threshold) are represented as an UpSet plot [21]. For each comparison (the top 30 are shown), the number of predicted splice-altering variants (vertical bars) and the tool(s) that identified these variants (solid dots) are shown. The horizontal bars show the total splice-altering variant identification rate of each tool, coloured as per legend

Back to article page