Predicting pathogenic protein variants

• 成果类型: Editorial Material
• 署名作者: Marsh, Joseph A.; Teichmann, Sarah A.
• 署名单位: University of Edinburgh; Wellcome Trust Sanger Institute; University of Cambridge
• 刊物名称: SCIENCE
• ISSN/ISSBN: 0036-13112
• DOI: 10.1126/science.adj8672
• 发表日期: 2023-09-22
• 页码: 1284-1285
• 关键词: tools

摘要

Many of the genetic mutations that cause disease in humans occur in protein-coding regions. Although the capacity to sequence DNA and identify these variants has substantially increased, the ability to interpret their effects remains limited. This problem is particularly acute for missense variants, which involve substitution of a single amino acid residue and make up the overwhelming majority of variants of uncertain significance (VUS), as classified by clinicians (1). On page 1303 of this issue, Cheng et al. (2) present AlphaMissense, a variant effect predictor (VEP) machinelearning algorithm that builds on the AlphaFold methodology for predicting protein structures from gene sequences (3). The authors demonstrate superior performance by AlphaMissense across multiple benchmarks compared with that of VEPs that are available now, which is likely to improve the interpretation of sequencing data and advance the role of computational predictions in the diagnosis of genetic disease.