Lack of African-relevant genomic resources have equated to Africans largely being excluded from the benefits of germline genetic testing. As such, disease-variant linked databases and mutation disease prediction guidelines remain European-biased. Irrespective of patient ancestry, predicting the disease impact of variants of unknown significance remains a challenge. While numerous in-silico variant pathogenicity prediction tools (VPPTs) have been developed, their applicability to African data is largely unknown. Using a multi-ethnic prostate cancer patient-matched whole genome variant resource, we compared 54 VPPTs across 145,291 known pathogenic or benign variants derived from 50 Southern African and 50 European men matched for advanced prostate cancer. Although sensitivity is lower for African data, we narrow VPPT selection to include both ancestrally shared and African-specific top-ranked performers. Using our guidelines, we close the ancestral disparity gap, increasing the prostate cancer prediction rate for our African-derived variants. Further disease-type-specific studies, with representation across the African diaspora, will be required to enable our findings to be translated into generalised and broadly African-inclusive germline testing criteria.
Results: Observing a 2.1- and 4.1-fold increase in the number of known and predicted rare pathogenic or benign variants, respectively, against a 1.6-fold decrease in the number of available interrogated variants in our European over African data, supports our hypothesis. Although sensitivity was significantly lower for our African data overall (0.66 vs 0.71, p = 9.86E-06), MetaSVM, CADD, Eigen-raw, BayesDel-noAF, phyloP100way-vertebrate and MVP outperformed irrespective of ancestry. Conversely, MutationTaster, DANN, LRT and GERP-RS were African-specific top performers, while MutationAssessor, PROVEAN, LIST-S2 and REVEL are European-specific. We narrowed the ancestral gap for potentially deleterious and oncogenic variant prediction in favour of our African data by 1.15- and 1.1-fold, respectively.