Reader Comments

Post a new comment on this article

Driver Mutations

Posted by RicDevor on 09 Apr 2012 at 18:10 GMT

As I understand the model developed here, you assume that any mutation carries a single, quantifiable value in oncogenesis. If that is true, then mutational events are members of a continuous distribution of effects. Is this not in conflict with the generally held view that there are mutations of major effect, so-called driver mutations, and mutations of little or no effect, so-called passenger mutations?

No competing interests declared.

RE: Driver Mutations

DonDai replied to RicDevor on 10 Apr 2012 at 13:47 GMT

We have taken an approach to expand the concept of driver and passenger mutations with an assumption that all mutations (and all genetic alterations) have a quantitative value for their effect on malignant transformation. Our approach seems to work well for quantitative analysis of complex interactions between genetic and environmental factors. The concept of driver mutation makes lots of sense if focused on the cause and molecular mechanism of human cancer. Knockout mouse models are often used to study gene function in oncogenesis. For instance, a mouse with knockout of TP53 gene often develops tumors, suggesting that alterations in TP53 are critical for oncogenesis. This seems to be confirmed in families with germline mutations of the TP53 gene. However, the same experiment can also be interpreted differently since the vast majority of cells in a mouse or a human with a p53 mutation are not cancerous. The counterargument is therefore that p53 mutations may play only a very small role, and other factors may play a much more important role. We have to assume that the number of "other factors" is fairly large in order to describe the cause of all human tumors. There are billions of possible genetic alterations in our genome. We believe that each should have a summary value for their effect on malignant transformation although the majority of them may be close to zero. A normal distribution of the quantitative mutation values will be able to explain why some cancers, along the tail of the normal distribution, may result from a single or a very small number of genetic alterations which have a large value but occur only very rarely. At the same time, the majority of human cancers, in the center of the normal distribution, harbor numerous genetic alterations with each of genetic alterations playing a small role. Recent findings from genome sequencing in pancreatic cancer (Jones, et al 2008. Science 321:1801) seem to suggest the involvement of a large number of mutations. The quantitative valuation of genetic alterations also sets the stage for a quantitative role of environmental factors, such as hormones, allowing mathematical expression to describe complex interactions of genetic and environmental factors. This valuation for all possible genetic alterations may not currently be quantifiable with current data and current evidence. But we can start with some genetic mutations which produce a much higher incidence in families with germline mutations. For instance, by comparing the lifetime risk of endometrial cancer in the general population versus in families with germline mutations, our model derived a numerical value for mutations in TP53, PTEN, MLH1, MSH2 and MSH6. The mutation with the most potent transforming effect in uterine epithelium is in the MSH6 gene, which has a small and negative value (suggesting a de-differentiating effect). In our model, the transforming value of a mutation is fixed. However, transformation of a cell into a cancer cell is determined by the combinational effect of many mutations and clonal selection induced by environmental pressure including hormone stimulation. The entire transformation process can be simulated along chronological time, with reference to a patient's age. The timing of mutational occurrence in cellular development also matters since a mutation in a proliferating cell has a much larger chance to induce a cancer than when the mutation occurs in a well-differentiated cell.

No competing interests declared.