Cancer Genomic Studies May Be Confounded by Mutational Heterogeneity

A major impetus for the initiation of large cancer sequencing projects has been the idea that large sample sizes will increase the power to identify genes mutated above the background mutation rate. However, as cancer genome study sample sizes have increased, the lists of significantly mutated genes have also grown. Lawrence and colleagues noted that these lists commonly include implausible candidate genes based on known biologic functions and are highly enriched for genes with specific genomic features, such as extremely long coding regions or introns. The authors hypothesized that current analytic methods identify so many specious genes because they fail to account for mutational heterogeneity that affects the background mutation rate. Using whole-exome or whole-genome sequencing data from over 3,000 matched tumor–normal pairs representing 27 different cancer types, they found that the mutation frequency varied by several orders of magnitude across patients with a given cancer type. This finding indicates that the assumption that cancers of a given type have a constant mutation rate is erroneous and could affect the accuracy of results. Another issue was that although individual tumor types tended to share a similar mutational spectrum, variability among individual samples suggested that current models of mutational processes used to calculate the background mutation frequency could be too simplistic. Finally, regional mutational heterogeneity appeared to be a key underlying cause of the identification of some suspect genes, as somatic mutation frequency was strongly correlated with low gene expression and late DNA replication. Taking these issues into account, the authors developed a method that would correct for patient- and gene-specific mutational heterogeneity and be capable of identifying much shorter lists of plausible significantly mutated genes. Accounting for mutational heterogeneity may therefore eliminate artifactual results in cancer genome studies and facilitate the identification and further validation of true cancer-associated genes.

Lawrence MS, Stojanov P, Polak P, Kryukov GV, Cibulskis K, Sivachenko A, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 2013;499:214–8.

Note: Research Watch is written by Cancer Discovery Science Writers. Readers are encouraged to consult the original articles for full details. For more Research Watch, visit Cancer Discovery online at http://CDnews.aacrjournals.org.