Small variant calling in the era of long-read sequencing.

Abstract

Long-read sequencing enables accurate analysis of complex and repetitive genomic regions but has distinct error profiles compared to short-read sequencing. Consequently, the tools for calling small variants - the predominant source of human genetic diversity - have shifted towards deep learning-based methods. This transition has reshaped variant calling across biological contexts, including germline, de novo, somatic and mosaic variants, and sequencing modalities for capturing both DNA and RNA at different resolutions. These deep learning frameworks are being developed in combination with alignment-based, assembly-based and pangenome-based strategies for small variant detection. In this Review, we chart the trajectory towards complete telomere-to-telomere profiling of small variants and consider the computational innovations needed to benchmark newly accessible repetitive regions and complex variants.