Signatures of Selection in Fusion Transcripts Resulting From Chromosomal Translocations in Human Cancer

Background: The recurrence and non-random distribution of translocation breakpoints in human tumors are usually attributed to local sequence features present in the vicinity of the breakpoints. However, it has also been suggested that functional constraints might contribute to delimit the position of translocation breakpoints within the genes involved, but a quantitative analysis of such contribution has been lacking. Methodology: We have analyzed two well-known signatures of functional selection, such as reading-frame compatibility and non-random combinations of protein domains, on an extensive dataset of fusion proteins resulting from chromosomal translocations in cancer. Conclusions: Our data provide strong experimental support for the concept that the position of translocation breakpoints in the genome of cancer cells is determined, to a large extent, by the need to combine certain protein domains and to keep an intact reading frame in fusion transcripts. Additionally, the information that we have assembled affords a global view of the oncogenic mechanisms and domain architectures that are used by fusion proteins. This can be used to assess the functional impact of novel chromosomal translocations and to predict the position of breakpoints in the genes involved.