Schmitz D, Ameur A, Johansson Å
Genome Res. - (-) - [2025-10-10; online 2025-10-10]
The T2T-CHM13 reference genome, released in March 2022, fills in the 8% of the human genome that was not resolved in GRCh38 and reconstructs large parts of the known genome. The more accurate and complete reference genome is expected to improve the quality of read mapping and variant calling. Even though whole-genome sequencing (WGS)-based approaches have become the gold standard in medical genetics, the extent of the benefits of the improved reference genome remains unclear. In this study, we aim to evaluate alignment and variant call performance with T2T-CHM13 as a reference using a cross-sectional Swedish cohort (SweGen) comprising 1000 individuals with short-read Illumina WGS data available. Remapping and variant calling is performed using the nf-core/sarek pipeline. T2T-CHM13 improves a wide range of mapping- and variant calling-related metrics, including a higher fraction of properly paired reads, lower mismatch rate, and more uniform coverage of coding regions. Moreover, the fraction of ambiguous alignments is higher, reflecting segmental duplications that were incorrectly collapsed in GRCh37 and GRCh38. In comparison to GRCh38, we identify 10 million additional variants in the cohort, including 5.5 million singletons, and observe an increased sensitivity for rare variants. SnpEff assigns impact ratings of moderate or high to 13% more variants in T2T-CHM13 than GRCh38. In summary, we conclude that T2T-CHM13 improves alignment metrics with higher alignment quality, better variant calling performance, and confidence, including for rare and deleterious variants. The T2T-CHM13 genome reference thus facilitates enhanced discovery of new disease-causing variation, benefiting, for example, rare-disease diagnostics.
NGI Uppsala (Uppsala Genome Center) [Collaborative]
National Genomics Infrastructure [Collaborative]
PubMed 40957660
DOI 10.1101/gr.279320.124
Crossref 10.1101/gr.279320.124
pii: gr.279320.124