Comprehensive benchmarking of somatic mutation detection by the SMaHT Network.

Short-read, long-read, duplex and single-cell benchmarking; donor-specific and pangenome-inferred diploid assemblies for somatic mutation calling

The Somatic Mosaicism across Human Tissues (SMaHT) Network conducted large-scale benchmarking experiments on cell lines and donor tissues to evaluate sequencing technologies, experimental approaches and computational methods for detecting different types of somatic mutations. It generated community resources with >1,000× short-read and 100-400× long-read data for nine analyzed samples. Donor-specific and pangenome-inferred diploid assemblies improved calling, extending mutation catalogs to challenging genomic regions. Six duplex technologies were benchmarked, single-cell sequencing resolved cell-type-specific mutational patterns and heterogeneity, and bulk, single-cell and duplex analyses proved complementary.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

This work primarily sets a methodological reference for somatic mosaicism research; its diagnostic reach is indirect, but it equips the detection of low-level mosaicism, a recognized cause of missed diagnoses in constitutional genetics. The depths used (>1,000×) are beyond routine clinical sequencing, and the abstract quantifies neither sensitivity nor allele-fraction thresholds. The announced roadmap still needs to be translated into recommendations applicable to patient samples.

Analysis by Dr Thibaut Benquey