Common fragile site profiling in epithelial and erythroid cells reveals that most recurrent cancer deletions lie in fragile sites hosting large genes.

Cancer genomes exhibit numerous deletions, some of which inactivate tumor suppressor genes and/or correspond to unstable genomic regions, notably common fragile sites (CFSs). However, 70%-80% of recurrent deletions cataloged in tumors remain unexplained. Recent findings that CFS setting is cell-type dependent prompted us to reevaluate the contribution of CFS to cancer deletions. By combining extensive CFS molecular mapping and a comprehensive analysis of CFS features, we show that the pool of CFSs for all human cell types consists of chromosome regions with genes over 300 kb long, and different subsets of these loci are committed to fragility in different cell types. Interestingly, we find that transcription of large genes does not dictate CFS fragility. We further demonstrate that, like CFSs, cancer deletions are significantly enriched in genes over 300 kb long. We now provide evidence that over 50% of recurrent cancer deletions originate from CFSs associated with large genes.