Underprediction of visibly complex chromosome aberrations by a recombinational-repair ('one-hit') model.

Purpose: Published low-LET FISH data were used to test two models of chromosome aberration production based on breakage-and-reunion or recombinational repair. Materials and methods: Randomness of DNA double strand break induction and misrejoining is analyzed comprehensively and adopted as a working hypothesis. Proximity effects are approximated by using interaction sites. Model results are calculated using CAS (chromosome aberration simulator) Monte Carlo computer software with two adjustable parameters. CAS can emulate the specifics of any experimental painting protocol, allowing very detailed tests of the models. Results: To reasonable approximation, breakage-and-reunion model predictions are consistent with low-LET FISH results, including two large, elaborate, one-paint data sets. An explicitly specified version of the recombinational-repair model severely underpredicts the frequency of the visibly complex aberration patterns most commonly observed with one-paint FISH, and is inconsistent with some observed multi-paint patterns. When high-dose effects (distortion and saturation) are taken into account quantitatively, a dose-response relation for apparently simple interchanges slightly favours the breakage-and-reunion model over the recombinational-repair model, despite being approximately linear over the dose range 2-6Gy. Conclusions: The random breakage-and-reunion model gives comprehensive baseline predictions that are sufficiently accurate for the organization of experimental results. The data speak against complex aberrations being formed by the random recombinational repair pathway discussed here. [ABSTRACT FROM AUTHOR]