Vancouver, BC, December 16, 2011--(T-Net)--Researchers in Simon Fraser University's Computational Biology lab have produced a new method of carrying out human genome comparisons that may help scientists and health care professionals predict the likelihood of conditions such as autism and mental retardation.
Using a new algorithm developed in SFU's School of Computing Science, the researchers, led by SFU professor Cenk Sahinalp, found that the strategy yields a much higher accuracy in detecting structural variations than the conventional methods used.
As a result, it is now possible to accurately detect the genetic variations among family members, especially between a child and the parents.
The findings imply much higher accuracy in pinpointing the potential genomic causes of certain conditions, such as autism, in children with healthy parents.
The algorithm is making international headlines within the genetics research community and is the cover story of the December 2011 issue of Genome Research.
Working with worldwide researchers on the 1000 Genomes Project, Sahinalp says, "This method compares genomes without first comparing them to a reference genome, so as to better predict genomic variations among closely related individuals, such as between a child and the parents.”
Sahinalp, who worked on the study with doctoral students Fereydoun Hormozdiari and Iman Hajirasouliha says scientists conventionally look for genome variations in two steps. First, short pieces from each individual genome are compared with an assembled reference genome, and its structural differences are identified.
The reference genome is a single genome sequenced and put together through the International Human Genome Project, completed in the early 2000s.
Next, the lists of structural variants in each genome are compared against each other. In this study, researchers moved away from the conventional approach to enable all genomes to be compared with the reference genome simultaneously, through a combinatorial optimization framework.
The focus of Sahinalp's lab is to combine discrete math, algorithms and molecular biology to predict genetic variation, especially large-scale variation, and to learn how these variations act as a precursor to developmental issues.
Photo credit: Computational Biology and Bioinformatics at SFU