ABSTRACT:

A sampling of the challenges and approaches involved with discovering cancer genes is presented. The following are also highlighted: detecting somantic genomic events; identifying driver genomic events; assessing functional impact; assessing statistical significance; and strategies for gaining confidence in algorithms for detecting cancer. In conclusion, one cannot readily determine the genetic events responsible for an individual's cancer simply by using massively parallel sequencing to sequence their tumor's genome, because the half dozen or so causal driver events are mixed with thousands of noncausal passenger events, millions of germline events, and tens of millions of instrumentation errors. Instead, many current approaches first identify the somatic events and then identify the subset of these that are driver events.