More powerful computers are allowing scientists and engineers to conduct simulations that grow more realistic each year. While companies are using these tools to slash the costs of producing everything from airliners to antibiotics, researchers in Houston are using them to refine their search for the genetic causes of disease.
In a new study published today in the journal PLoS Genetics, statisticians and genetic epidemiologists from Rice University and The University of Texas M. D. Anderson Cancer Center used computer simulations to trace genetic changes over thousands of generations in a simulated population of hundreds of thousands of people.
The goal: find out whether the tools that statistical geneticists use to pinpoint disease genes are up to the task of identifying multiple genes that cause complex diseases like cancer.
“In a real population, you never have the complete genetic picture, particularly for complex diseases where more than one gene is implicated and where environmental factors play a role,” said lead author Bo Peng of M. D. Anderson. “If we only see the people who get sick, we can never be sure how many people with the disease variant of the gene avoided getting sick. And there’s always the question about how many people got the disease even though they didn’t carry the variant.”
In order to simulate the evolution of complex human diseases, Peng developed a computer program called simuPOP that generates genetic profiles for large multi-generation populations. The program, which Peng developed during his doctoral studies at Rice, allows researchers to sample individuals from a simulated population and test whether statistical methods are up to the task of accurately identifying genes that interact to cause complex diseases.
“Though they have much in common, the disciplines of statistical genetics, population genetics, molecular genetics and genetic epidemiology have traditionally used their own tools and techniques,” said co-author Marek Kimmel, professor in the Statistics Department at Rice. “simuPOP is one of the first examples of a new paradigm where the tools of the various disciplines are being used in concert to create a clearer picture of genetic health effects.”
“Complex diseases like hypertension and cancer are usually caused by multiple disease-susceptibility genes, environmental factors and interactions between environmental and genetic factors,” said co-author Christopher Amos, professor of epidemiology at M. D. Anderson. “In the current study, we show that our method of simulating populations as they move forward in time, over multiple generations, is a practical and useful approach for simulating complex diseases.”
Peng said the latest findings are preliminary but they confirm that known statistical genetic methods are limited in their ability to accurately identify the genetic interactions implicated in complex diseases. Peng said the findings are useful because they identify which methods work best with particular types of populations. He said simuPOP could be useful in developing and testing new methods for gene mapping, and he’s already gotten interest from more than 20 research groups that are interested in using the program.