11 genes responsible for Type 2 Diabetes identified thanks to...mathematicians

While it may seem an odd realm for mathematicians to be involved in, a team of mathematicians from Michigan Technological University have recently tried their hands at genetics. The team has built a series of programs which allow for the determination of specific gene combinations involved in severe and hard to treat diseases, particularly Type 2 Diabetes. Quiying Sha, an assistant professor at MTU, explains, “With chronic, complex diseases like Parkinson's, diabetes and ALS [Lou Gehrig's disease], multiple genes are involved [and] you need a powerful test.”

Previous attempts at determining the genes responsible for diseases such as Type 2 Diabetes have been an overwhelming task due to the sheer number of calculations required for possible gene combinations of the roughly 500 000 genes in the human genome. As such the team of mathematicians has developed not one but two powerful programs to handle these calculations, known as the Ensemble Learning Approach (ELA). The first program is designed to trace possible genes responsible for Type 2 Diabetes back through generations, much further back than previous analysis, allowing for the majority of genes to be ignored, thus greatly narrowing the number of possible combinations responsible for the disease.

With the field of possibilities now suitably thinned the genes are fed into the second program, this one a purely statistical system calculating possible combinations. After using the ELA to test genes from 1000 people, half without Type 2 Diabetes and the remaining half sufferers of Type 2 Diabetes, the ELA returned to the team 11 variations, known as single nucleotide polymorphisms (SNPs), within human genes that, separately or in various combinations, have a high probability of causing Type 2 Diabetes.

While the discovery of these 11 SNPs is a large achievement it is overshadowed by the future prospects for the Ensemble Learning Approach. With the programs now operational analysis is “relatively simple” according to Sha, and with the necessary data sets any genetic disease may be analysed for the genes responsible.

Jared Miles (42024211)

Article: http://www.sciencedaily.com/releases/2009/10/091006121115.htm

Picture: http://repairstemcell.files.wordpress.com/2009/02/diabetes_0