Understanding Specific Genetic Mutations that Drive Cancer

When a patient asks, “Why me?” after learning they have cancer, their physician’s response likely includes a review of the patient’s risk factors.

“Historically, risk factors are nearly the only answer that science or medicine has been able to give to cancer patients who ask, ‘why me?’– or to public health officials who ask, ’why us?’” said Jeffrey Townsend, Yale’s Elihu Professor of Biostatistics and professor of ecology and evolutionary biology.

Townsend and his colleagues’ innovation is to answer the question, “why me?” for individual patients. A new open-source software package called cancereffectsizeR developed by the Townsend Lab greatly improves the ability of data scientists and clinical analysts to learn about the specific genetic mutations that drive cancer.

The software calculates the effect sizes of single-nucleotide variants in cancer, quantifying their effect on the ability of the cancer cells to proliferate and survive in humans. The unique algorithm does so by organizing somatic variant data, facilitating mutational signature analysis, and calculating site-specific mutation rates. The software can quantify cancer effects at specific stages of cancer evolution, and in the contexts of other key mutations, using the somatic evolutionary information in their tumor DNA. Then it relates those effects to machine-learned signatures of specific sources of mutations.

“Our answer tells individual patients what proportion of the origination of their cancer can be attributed to known mutagenic causes such as aging, smoking, ultraviolet light, and haloalkane exposure,” Townsend said. “Local populations or professions that suffer from inordinately high levels of cancer may also be able to use the findings to discover instances of exposure to carcinogenic substances.”

Townsend added, “Our results provide molecular validation of well-known correlative findings from the epidemiological literature.” For example, he explained that melanomas are often largely attributable to preventable, exogenous exposure to ultraviolet light, and lung cancers are often largely attributable to preventable, exogenous exposure to tobacco. On the other hand, gliomas and prostate adenocarcinoma tumors that are largely attributable to endogenous processes associated with aging are not subject to public health efforts at prevention.

“Our innovation rapidly directs research efforts toward these most important causes, and toward effective public health prioritization, by informing all of us of what actions we can take to prevent cancer,” Townsend said. “If you want to know what caused the mutations that drive your cancer, our innovation enables its calculation. Next steps include applying this method to specific populations with elevated cancer incidence to determine directly whether specific mutagenic sources are responsible – a major boon to the prevention of the next public health threat.”