DCS Summer 2012 Research Awards -- Project Description

Identifying Well-Matched Controls for Medical Sequencing

Faculty name: Mike Brudno
Research area: Bioinformatics, Artificial Intelligence
Campus address: PT 286C
Campus phone: 416-978-2589
Email address: brudno [at] dgp.toronto.edu
Number of students: 1
Skills required:
  • Extensive knowledge of biology is not required, but a good understanding of Algorithms (CSC 373 or at least 263), and some exposure to statistics are necessary.

Brief project description:

When medical scientists decipher the genomes of multiple patients with a certain disorder, they often compare the frequencies of observed differences to those in healthy, or "control" individuals. If a certain genomic variant appears at a significantly different frequency in cases versus controls, it is a good candidate for follow-up study.

The accuracy of such comparisons, called Genome-Wide Association Studies, depends on having a set of controls that is well matched (i.e. coming from a similar ethnic background). While many projects have ascertained the genomes of thousands of healthy individuals, identifying a subset of these that matches a set of patients with a disorder is an open problem.

In this project you will be responsible for developing and implementing an algorithm, that given a very large set of healthy individuals, identifies a smaller subset of them that most closely matches a smaller disease cohort.

Back to the index.