Pawel F. Przytycki will present his pre-FPO Thursday February 15th, 2018 at 11am in LSI room 253. The members of his committee are are Mona Singh (Advisor), Olga Troyanskaya, Ben Raphael, Josh Akey (Ecology & Evolutionary Biology) , and Yibin Kang (Mol Bio). All are welcome to attend. Abstract and title follow below. Algorithms for deciphering cancer genomes: from differential mutation to differential allele specific expression Large-scale cancer genome sequencing consortia have provided a huge influx of somatic mutation data across large cohorts of patients. Understanding how these observed genetic alterations give rise to specific cancer phenotypes represents a major aim of cancer genomics. A common approach is to test which genes and genomic locations tend to accumulate more changes across tumors. However, using mutation frequency alone is not sufficient due to gene-specific characteristics such as length, replication timing, and expression, which all play a role in any given gene’s propensity for acquiring mutations. In this talk I will present two methods for utilizing natural variation as a background for understanding cancer. First, I will introduce differential mutation analysis, a framework for uncovering cancer genes that compares the mutational profiles of genes across cancer genomes with natural germline variation across healthy individuals. DiffMut, a fast and simple approach for differential mutational analysis outperforms considerably more sophisticated approaches at discovering cancer genes across 24 cancer types by utilizing natural variation data from the 1000 Genomes Project. Second, I will present a method for detecting functional noncoding somatic mutations using allele specific expression. While much of the focus of previous cancer research has been on coding mutations, the vast majority of mutations occur in noncoding regions. The changes caused by these mutations can be difficult to detect and to date only a few driver mutations have been identified. When noncoding mutations lead to regulatory changes, the effect is often allele specific because the mutation only occurs in one copy of the individual's DNA. However, in the case somatic mutations, changes only occur in cancer cells which only form a fraction of the tumor sample. Furthermore, many genes already display allele specific expression in a normal state. The method I have developed links noncoding mutations to allele specific expression by looking at both allele expression in the tumor and a matched normal using a series of models with increasing sensitivity based on different assumptions of tumor heterogeneity and differential expression. I validate these models on simulations and show that they are able to detect allele specific expression resulting from nonsense mutations. When applied to noncoding mutations in breast tumors, the models are able to uncover mutations upstream of genes that appear to exhibit differential allele specific expression.