Dissecting in silico Mutation Prediction of Variants in African Genomes: Challenges and Perspectives

Abstract

Genomic medicine is set to drastically improve clinical care globally due to high throughput technologies which enable speedy <i>in silico</i> detection and analysis of clinically relevant mutations. However, the variability in the <i>in silico</i> prediction methods and categorization of functionally relevant genetic variants can pose specific challenges in some populations. <i>In silico</i> mutation prediction tools could lead to high rates of false positive/negative results, particularly in African genomes that harbor the highest genetic diversity and that are disproportionately underrepresented in public databases and reference panels. These issues are particularly relevant with the recent increase in initiatives, such as the Human Heredity and Health (H3Africa), that are generating huge amounts of genomic sequence data in the absence of policies to guide genomic researchers to return results of variants in so-called actionable genes to research participants. This report (i) provides an inventory of publicly available Whole Exome/Genome data from Africa which could help improve reference panels and explore the frequency of pathogenic variants in actionable genes and related challenges, (ii) reviews available <i>in silico</i> prediction mutation tools and the criteria for categorization of pathogenicity of novel variants, and (iii) proposes recommendations for analyzing pathogenic variants in African genomes for their use in research and clinical practice. In conclusion, this work proposes criteria to define mutation pathogenicity and actionability in human genetic research and clinical practice in Africa and recommends setting up an African expert panel to oversee the proposed criteria.