Smaller software changes are not announced here. A summary of the three-weekly release changes can be found here. For the full list of our daily code changes head to our GitHub page. Lastly, see our credits page for acknowledgments of the data we host.
We are pleased to announce the release of the ClinPred pathogenicity score track for -hg19 and -hg38. +hg19 and +hg38. ClinPred is a machine-learning predictor of pathogenicity for nonsynonymous (missense) single-nucleotide variants, combining existing pathogenicity scores with population allele frequency from gnomAD. It was trained on confidently annotated disease-causing and benign variants from ClinVar. Pre-computed scores are provided for all possible human missense variants in the exome.
Scores range from 0 to 1, with higher values indicating a greater predicted likelihood that a variant is disease-relevant. The authors recommend a score of ≥ 0.5 as evidence of pathogenicity. As with any pathogenicity prediction score, ClinPred is intended as supporting evidence rather than a stand-alone classifier.