Math Biology Seminar: Hisashi Ohtsuki

A traditional population-genetics approach studies geneaologies in a population of a fixed size, which forms the basis of several spectral theories of finite samples. In contrast, a population of tumor cells typically experiences an exponential growth phase in its initial progression, which is far from constant population size. In this work, I develop two different numerical procedures, one of which is based on forward-in-time and the other is based on backward-in-time treatment, to derive allele frequency spectrum in such exponentially growing cancer cell populations. We find significance bias toward singletons both analytically and numerically, which reflects the fact that most observed mutations have recent origins in a growing population. This work was done in collaboration with Prof. Hideki Innan