henhaobio
Sign in

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (20 Aug 2026) · CC BY 4.0

Duplex sequencing of a Li-Fraumeni family links germline TP53 p.R181H to higher neutral-region mutagenesis and weaker blood TP53 selection, plus recurrent p.R248Q/W hits across proband tissues.

Li-Fraumeni syndrome (LFS) carriers of germline TP53 variants develop early cancers, but it is poorly understood how one altered TP53 copy shapes mutation and clonal selection in normal tissue before cancer, because LFS is rare and standard sequencing misses tiny somatic clones. This work applies error-corrected duplex sequencing to blood from a p.R181H family and unrelated controls, plus autopsy tissues from the affected proband, to measure those pre-cancer dynamics.

Key results

  1. SuggestedGermline p.R181H was associated with higher mutation frequency (MF) in a neutral 20-region mutagenesis panel. Coding MF in clonal hematopoiesis (CHIP) genes tracked age and chemotherapy but not carrier status, and CHIP clone burden tracked age only.

    Multivariate Gaussian generalized linear models (covariates: age, LFS, chemotherapy) on blood at mean duplex depth 15,683× (12,606× - 18,282×). Pooling all individuals, tissues and panels, the study found 1,992 CHIP coding and 6,487 mutagenesis-panel mutations. Mutagenesis MF showed no chemotherapy effect. TP53 non-coding MF was linked to p.R181H; TP53 coding MF tracked only chemotherapy. · Ultra-deep sequencing of a family carrying…; Germline p.R181H does not affect most CHIP…; Chemotherapy increases TP53 clonal expansions in…; Regression analysis

  2. SuggestedSelection on CHIP genes was mostly independent of carrier status (positive for DNMT3A and TET2, negative for GATA2). TP53 was the exception: it lacked positive selection in carriers, and untreated carriers had TP53 dN/dS below 1 once chemotherapy-exposed people were analyzed separately.

    TP53 was the only gene whose bootstrapped (5,000 resamples) dN/dS interquartile ranges did not overlap between groups. In untreated carriers LFS02/LFS03, 4/7 (57%) TP53 coding mutations were non-synonymous vs 18/22 (82%) in non-LFS individuals (not significant). Chemotherapy raised TP53 dN/dS in both groups. · Germline p.R181H does not affect most CHIP…; Chemotherapy increases TP53 clonal expansions in…; Testing for selection

  3. ShownAcross the proband's autopsy samples, TP53 coding MF was elevated in sun-exposed skin, blood, normal esophagus and both chronically inflamed gastric samples. All tumors had lost the wild-type TP53 allele and showed no excess coding MF.

    Sun-exposed skin had more than twice the TP53 mutations of the next sample, many of them UV-type CC>TT dinucleotides (SBS7a/b/d). Primary and metastatic samples had p.R181H variant allele frequency >0.5. SBS17a/b appeared in esophageal tumors, metastases and one normal esophagus. Four samples at 2,118-7,376× depth were excluded from MF. · Ultra-deep sequencing of a family carrying…; Mutation frequency and signatures vary across…; Mutational spectrum and signature analysis

  4. ShownA TP53 codon 248 substitution (p.R248Q or p.R248W, both inactivating and dominant negative) recurred at low frequency in most of the proband's samples. TP53 mutations were enriched in the DNA-binding domain and had high AlphaMissense scores, consistent with convergent selection.

    18/28 tissue samples carried a codon 248 mutation, and 9/19 were seen in >1 duplex read. The largest buffy-coat clone had frequency 0.0005, and most blood mutations were not shared across tissues, arguing against blood carryover. In cohort blood, 75% of TP53 SNVs fell in the DNA-binding domain vs 46% expected (binomial test). · Chemotherapy increases TP53 clonal expansions in…; TP53 R248Q/W is a mutational hotspot recurrently…

  5. ShownAmong somatic TP53 mutations that could be phased against the germline site, most arose on the chromosome copy without p.R181H, pointing to selection for second hits on the wild-type allele.

    Duplex reads spanning the germline site (hg38 chr17:7675070; candidate variants within 150 bp) phased 29 mutations across LFS01, LFS02 and LFS03. Of these, 25/29 lay on the non-carrier copy (binomial test, 0.0001), and all spanning reads agreed on phase. · Allelic specificity but not tissue-context…; Testing for selection

Methods, data and reproducibility

Duplex (double-strand consensus) sequencing with TwinStrand hybrid-capture panels (29-gene CHIP/AML panel, 20-region neutral mutagenesis panel, TP53 panel) on blood and autopsy tissues, then regression of mutation frequency, bootstrapped dN/dS, mSigHDP signature extraction, AMSD spectrum comparison, and read-level phasing against p.R181H.

  • Proband LFS01 (germline p.R181H; died of metastatic esophageal adenocarcinoma at 34): hematopoietic samples (buffy coat, plasma, whole blood, bone marrow) · human
  • LFS01 autopsy non-cancer tissues (e.g., esophagus, stomach, liver, colon, kidney, thyroid, testis, skin); all TP53 panel, a subset also mutagenesis panel · human · n = 18 non-cancerous samples spanning 16 tissues
  • LFS01 tumors: primary esophageal samples plus metastases from liver, lung and mediastinum · human · n = 6 cancerous samples
  • PBMCs from relatives: two p.R181H carriers without known cancer (LFS02, LFS03) and one non-carrier relative (REL01) · human · n = three family members
  • PBMCs from unrelated healthy controls CON01-CON07 aged 25 to 76; CON07 previously received CHOP chemotherapy · human · n = 7 unrelated control donors
  • COSMIC v102 (GRCh38) TP53 codon-level mutation counts across tumor types, used as a hotspot reference · human
Methods & tools
Duplex sequencing (TwinStrand V1 kits) on Illumina NovaSeq 6000, TwinStrand AML-29 Assay (CHIP panel) and Human Mutagenesis Assay, TwinStrand DNAnexus pipeline 3.20.1 (VarDict-based caller), Vcf2Maf v1.6.22 with Ensembl VEP, AlphaMissense, mSigHDP (hierarchical Dirichlet process signature extraction), SigProfilerMatrixGenerator v1.3.3, SigProfilerExtractor v1.1.21 with COSMIC v3.4 signatures, AMSD (aggregate mutation spectrum distance) v0.3.0, Gene-level dN/dS with bootstrap resampling, pysam v0.22.1 (read-level phasing), COSMIC v102; UCSC RepeatMasker
Software & environment
TwinStrand software 3.20.1 on DNAnexus; Vcf2Maf v1.6.22; Ensembl VEP 112 and 115; R 4.5.1 (RStudio 2025.09.1); Python 3.12.3 with pysam v0.22.1; SigProfilerMatrixGenerator v1.3.3; SigProfilerExtractor v1.1.21; AMSD v0.3.0
Reproducibility gaps
Raw sequencing data are not yet deposited; the authors say they will go to NCBI BioProject before publication. Variant calling uses the TwinStrand pipeline hosted on DNAnexus. R package versions appear only in a supplemental table.
Limitations (6)
  • There were no normal-tissue controls, and no published study has sequenced matched tissues at comparable depth, so the proband's tissue mutation levels have no reference. · stated by the authors
  • Multi-tissue data come from one proband who received chemotherapy (including 5-fluorouracil), which confounds selection estimates. Because SBS17b also overlaps with 5-fluorouracil exposure, the drivers of esophageal mutagenesis cannot be cleanly separated. · stated by the authors
  • Tissues were sequenced only for TP53 (plus the mutagenesis panel in a subset), a narrower footprint than other studies, though at much greater depth (15,000x versus <1000x). Tissue dN/dS had to be pooled into broad categories, which may hide site-specific selection. · stated by the authors
  • Only one germline variant was studied. Because p.R181H has reduced penetrance, an atypical cancer spectrum and later onset, the findings may not generalize to other LFS variants. · stated by the authors
  • Carrier effects rest on three related p.R181H carriers (two untreated) and a small control set, with one blood sample per person in the regressions, so familial background and age cannot be fully separated from the variant. · noted in this summary
  • The proband's blood comparator was buffy coat, while all other subjects contributed PBMCs. This difference in sample type could shift CHIP and mutagenesis-panel mutation frequencies independently of carrier status. · noted in this summary

Read this if you study clonal hematopoiesis, TP53 second-hit dynamics, or somatic evolution in the normal tissues of cancer predisposition carriers and want ultra-deep duplex sequencing data from a Li-Fraumeni family.

Citations
0

Abstract

Li-Fraumeni Syndrome (LFS) is caused by germline pathogenic variants in TP53 which predispose carriers to early onset cancer across multiple tissues. While genomically profiling those cancers has revealed factors contributing to their formation, little is understood about how LFS impacts clonal evolution in healthy tissues preceding cancer. Here, we use ultra-deep duplex sequencing (mean [~]15,000x depth) to investigate somatic mutation and selection in a family carrying the germline TP53 p.R181H pathogenic variant and a cohort of non-carrier controls. In blood samples, the germline variant is associated with more mutations in a panel designed to capture genomewide mutagenesis, and with reduced positive selection on somatic TP53 mutations, despite confounding by chemotherapy treatment in one individual. DNMT3A and TET2 mutations appear positively selected and GATA2 mutations negatively selected across the cohort, independent of the p.R181H status. Extensive multi-tissue sampling of 22 non-cancerous and 6 cancerous samples was also performed at autopsy in one individual with LFS who succumbed to esophageal cancer. Cross-tissue analysis reveals excess mutations in sun-exposed skin, esophagus and chronically-inflamed stomach tissue, and concordant mutations in the p.R248 hotspot of TP53 across most (18/28) tissue samples. Most somatic TP53 mutations in LFS that can be assessed for phase arose on the chromosomal copy lacking the p.R181H variant. Our study reveals how the germline p.R181H variant reshapes baseline somatic mutation and selection in normal tissues and highlights the importance of understanding early somatic evolution in LFS prior to cancer development and treatment.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISMacGregor '26Montellier '25Kobar '24Montellier '24Shin '19LI '19
71% semantic matchGenetic & Genomic Medicine
Hamish A J MacGregor, Jamie R. Blundell, Douglas F. Easton·The American Journal of Human Genetics·Aug 17, 2026· 1 citations