In cells proliferating mitotically, there are two general strategies for DSB repair: end joining and homology-directed repair (HR). Our interests are in HR. In contrast to HR that occurs during meiotic recombination, HR in proliferating cells rarely produces crossovers. Rather, non-crossover products are produced, though the broken chromatid may experience gene conversion. There are several different "pathways" for HR. In our view, the most common of these is probably synthesis-dependent strand annealing (SDSA):

The SDSA model

• Resection of the 5' ends, leaving 3'-ended single-stranded tails.
• Rad51-mediated homology search, followed by strand invasion, in which the resected strand invades a homologous duplex (usually the sister chromatid or homologous chromosome), displacing the strand of same polarity to generate a D-loop structure.
• Repair DNA synthesis.
• Dissociation of the nascent strand. We beleive that repair synthesis is not highly processive, but extends the invading strand by perhaps several hundred nucleotides before dissociation. In this figure, dissociation is drawn as if it occurs after synthesis is completed; it is also possible that dissociation occurs during synthesis, so that a migrating D-loop travels down the template.
• Annealing of the dissociated strand to the other resected end of the DSB.
• Completion of repair requires filling of gaps and/or trimming of overhangs, repair of mismatches if the template was not identical to the broken chromatid, and ligation (these steps are not shown).

For many years, SDSA got little respect, in part because of strong support for the DHJ model during meiotic recombination, and in part because most assays cannot distinguish between SDSA and other models. We have develope a gap repair assay for SDSA. In particular, we have been examining the role of DmBLM, the Drosophila ortholog of the Bloom syndrome helicase, in promoting SDSA as a mechanism to prevent mitotic crossing over.