Back

Timing the onset of homologous recombination deficiency before cancer diagnosis

Andreopoulos, M.; Niu, M.; Zhang, Y.; Viswanadham, V. V.; Gulhan, D. C.; Jin, H.; Batalini, F.; Wulf, G.; Zong, C.; Park, P. J.; Glodzik, D.

2025-10-19 genomics
10.1101/2025.10.17.683157 bioRxiv
Show abstract

Mutations in BRCA1 and BRCA2, whether inherited or somatically acquired, cause homologous recombination deficiency (HRD) in tumor cells. However, the precise timing of HRD onset in the tumor lineage is unknown. Under the haploinsufficiency model of BRCA1/2, HRD-associated mutagenesis could begin as early as fertilization, whereas the prevailing two-hit tumor suppressor model places HRD onset upon bi-allelic loss, the exact timing of which remains undetermined. Here, we analyzed mutational signatures in 118 breast and ovarian cancer genomes exhibiting HRD to estimate the onset of HRD-driven mutagenesis prior to cancer diagnosis. Using HRDTimer, a novel algorithm, we estimate that HRD arises in tumor precursor cells at 37% of SBS1-based molecular time--corresponding to 8.6 years (IQR 7.8-9.7) prior to diagnosis in triple-negative breast cancers, and 16.0 years (IQR 14.2-18.5) in ER-positive breast cancers. Bulk sequencing reveals an accelerated accumulation of SBS1 mutations following neoplastic transformation, influencing the age estimate of HRD onset. Single-cell duplex sequencing confirms this pattern and further shows that nearly all non-tumor cells lack the HRD signature, indicating that HRD is rare in pre-malignant cells, even in patients with inherited BRCA1/2 mutations. Our analysis provides an estimate of the time window during which HRD mutagenesis emerges years before diagnosis, offering quantitative insight into when detection and potential interception may be possible.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.