Houtain et al. (Science Advances, 2026) pair a ~75β135 generation mutation-accumulation (ME) experiment (22 lines) with acute genotoxic treatments (30-day desiccation; 100β500 Gy proton irradiation) to trace loss of heterozygosity (LOH) in the asexual bdelloid rotifer Adineta vaga. In ME lines, 85.2% of LOH clusters are small (<100 sites, <~5 kb), consistent with gene conversion during nonreductional meiosis I; large clusters (>100 sites, 14.8%) carry 85.8% of total LOH, with 19.4% of them flagged as putative deletions . The headline result: after 500 Gy, large hemizygous deletions (up to ~400 kb) show allele-frequency shifts and gradual coverage recovery over ~6 months of culture, and two smallest deletions fully restored to ~1.0 coverage β supporting a new concept, break-induced homologous extension repair (BIHER), operating across generations . Karyotyping found more DAPI-stained bodies in 500-Gy descendants (computer-vision means 14.67 vs 9.73 controls), consistent with inherited chromosome fragments . These findings extend the authors' earlier demonstration of delayed, HR-based germline repair during nonreductional meiosis .
Strengths: The LOH/coverage framework is quantitative, dose-response LOH to proton IR was significant (OLS P=1.18e-3; Tukey 500 Gy vs 0 Gy P=0.0042), simulations bound the interpretation space, and Nanopore reads lacked religation intermediates, arguing fragments stay unsealed . Data and code are public (BioProjects PRJNA680543, PRJNA929432, PRJNA930383; github.com/AntoineHo/genome) .
Weaknesses and blindspots: (1) BIHER's molecular mechanism is inferred, not demonstrated: coverage recovery could fit several scenarios (longer repair tracts, higher repair probability, or sporadic full repair); the authors admit the specific mechanism is "difficult to disentangle." (2) Holocentricity is proposed, not proven β Hi-C merely failed to show discrete centromeres. (3) The 6-month Nanopore resequencing covers only one clone (P500 C18 E3, 8 deletions); generalization across lines is limited. (4) Coverage/AF profiles partly reflect subclonal heterogeneity, so bulk sequencing averages over individuals; single-cell or targeted-break tracking would be decisive (the authors note this is ongoing). (5) Bottlenecks and untracked generations during desiccation-recovery may inflate LOH-rate estimates. (6) The claimed ~88β440 DSBs/genome per dose is theoretical, not measured per-sample.
What would falsify BIHER: small clusters dominating over large ones in ME lines (not observed: 85.2% are small), IR-induced deletions showing no temporal coverage gain, or religated intermediates detected in long reads. Conversely, targeted Cas9-induced breaks followed over generations with temporal resolution would strongly test BIHER's progressive repair predictions.
Context: The finding reframes bdelloid "evolutionary scandal" asexuality: recombinational repair without reductional meiosis supplies homologous exchange benefits while preserving heterozygosity, complementing evidence of long-term asexuality from divergent gene copies and the somatic-vs-germline division of repair labor via NHEJ/BER early and HR delayed in oocytes .
Know what changed, what holds up, and what remains uncertain. Every Friday. No ads.