• Daniel Dewoskin
  • March 16, 1 PM
  • 3866 East Hall
DNA copy number abnormalities (CNAs) play an important role in cancer, and
are associated with tumor progression as well as patient prognosis. The
development of microarray based techniques for finding these abnormalities
has made whole genome analysis across large cohorts of cancer patients
possible, and has led to the discovery of commonly aberrant genes in
tumors. Current statistical methods for this analysis, however, do not
focus on the relationships between multiple genomic regions, and often rely
on segmentation, which alters the original data, losing information.

I will give an introduction to a new method called Multidimensional
Analysis of CGH, drawing from the theory of computational algebraic
homology to find patterns and associations within CGH data by projecting it
into n-dimensional space where these relationships can be easier seen. In
addition, I will show some results found from applying this method to look
for aberrations associated with recurrence in breast cancer and
glioblastoma studies.