A Commentary on Lawson ARJ, et al. Somatic mutation and selection at the population scale. Nature. 2025;647(8089):411–420. <a href="https://doi.org/10.1038/s41586-025-09584-w">https://doi.org/10.1038/s41586-025-09584-w</a>. Design A cross-sectional observational cohort study using advanced genomic sequencing to map somatic mutations and clonal selection in normal human tissues at the population level<CitationRef CitationID="CR1">1</CitationRef>. Participants The authors used targeted NanoSeq, a duplex sequencing method achieving exceptionally low error rates (fewer than five errors per billion base pairs). The technique was applied to targeted capture of a 239-gene panel (0.9 Mb) in 1042 non-invasive buccal swab samples (median donor age 68 years, 79% female, 37% smokers) and 371 blood samples from the TwinsUK registry. Data analysis Mutation calling required strict duplex consensus filters. Gene-level selection was quantified using dNdScv, while site-level dN/dS analyses were used to identify recurrent hotspots and selected sites. Mutational signatures were deconvolved with sigfit, and multivariate mixed-effects regression modelled associations with age, smoking (pack-years), alcohol (drink-years), missing teeth, and other covariates. Results In the oral epithelium, single nucleotide variants (SNVs) accumulated at 18.0 per cell per year and indels (insertions or deletions in the genome) at 2.0 per cell per year. Across all 1,042 participants, the authors detected 341,682 somatic mutations, including approximately 62,000 estimated driver mutations across 46 positively selected genes (e.g., NOTCH1 mutant cells reaching approximately 10% by ages 65–85 years, TP53 ~ 3%). Nine genes showed negative selection. Two dominant mutational signatures emerged: a clock-like signature A (resembling SBS1 and SBS5) and an alcohol associated signature B (resembling SBS16). The authors reported that smoking and alcohol were associated with higher mutation burdens, and that some effects on clonal selection were inferred from regression analyses; missing teeth (a marker of poor oral health) were independently associated with higher overall driver density. Driver frequencies for NOTCH1 were similar in normal epithelium and head and neck squamous cell carcinoma (HNSCC), whereas TP53 was enriched in cancer. Conclusions The authors concluded that optimised NanoSeq enables population-scale mutational epidemiology in polyclonal tissues, revealing an extraordinarily rich landscape of positive and negative selection in normal oral epithelium. Environmental exposures shape mutagenesis and may also influence clonal selection, supporting a plausible mechanistic link between smoking, alcohol, and oral cancer risk.

Uncovering the hidden mutational landscape of the oral epithelium

期刊论文
出版模式：

开放获取
发表日期：

2026年5月21日

查看全文

Manas Dave

摘要

A Commentary on

Lawson ARJ, et al.

Somatic mutation and selection at the population scale. Nature. 2025;647(8089):411–420. https://doi.org/10.1038/s41586-025-09584-w.

Design

A cross-sectional observational cohort study using advanced genomic sequencing to map somatic mutations and clonal selection in normal human tissues at the population level¹.

Participants

The authors used targeted NanoSeq, a duplex sequencing method achieving exceptionally low error rates (fewer than five errors per billion base pairs). The technique was applied to targeted capture of a 239-gene panel (0.9 Mb) in 1042 non-invasive buccal swab samples (median donor age 68 years, 79% female, 37% smokers) and 371 blood samples from the TwinsUK registry.

Data analysis

Mutation calling required strict duplex consensus filters. Gene-level selection was quantified using dNdScv, while site-level dN/dS analyses were used to identify recurrent hotspots and selected sites. Mutational signatures were deconvolved with sigfit, and multivariate mixed-effects regression modelled associations with age, smoking (pack-years), alcohol (drink-years), missing teeth, and other covariates.

Results

In the oral epithelium, single nucleotide variants (SNVs) accumulated at 18.0 per cell per year and indels (insertions or deletions in the genome) at 2.0 per cell per year. Across all 1,042 participants, the authors detected 341,682 somatic mutations, including approximately 62,000 estimated driver mutations across 46 positively selected genes (e.g., NOTCH1 mutant cells reaching approximately 10% by ages 65–85 years, TP53 ~ 3%). Nine genes showed negative selection. Two dominant mutational signatures emerged: a clock-like signature A (resembling SBS1 and SBS5) and an alcohol associated signature B (resembling SBS16). The authors reported that smoking and alcohol were associated with higher mutation burdens, and that some effects on clonal selection were inferred from regression analyses; missing teeth (a marker of poor oral health) were independently associated with higher overall driver density. Driver frequencies for NOTCH1 were similar in normal epithelium and head and neck squamous cell carcinoma (HNSCC), whereas TP53 was enriched in cancer.

Conclusions

The authors concluded that optimised NanoSeq enables population-scale mutational epidemiology in polyclonal tissues, revealing an extraordinarily rich landscape of positive and negative selection in normal oral epithelium. Environmental exposures shape mutagenesis and may also influence clonal selection, supporting a plausible mechanistic link between smoking, alcohol, and oral cancer risk.