The Causal Role of Skin Pigmentation in Melanoma: A Machine-Learning Based Gene Prioritization Study
|Hensin Tsao, MD, PhD
|Medical Student Award
|Massachusetts General Hospital
|Randy Lomax Memorial Medical Student Award
Pigmentation, or the coloring of the skin, hair and mucous membranes, has been shown to influence melanoma risk. Most, if not all traits and diseases, have a genetic component that influences development or susceptibility. These genetic components, or loci, can be determined using genome-wide association studies (GWAS). GWAS studies associate millions of relatively common genetic modifications between a healthy population and a population containing a trait or disease of interest. The assignment and interpretation of these genetic modifications has also proved challenging. In this study, we aim to utilize a series of computational approaches to determine whether GWAS associations between pigmentation and melanoma are truly causal rather than due to coincidence. We then use a machine learning algorithm to identify and assign a likelihood of gene involvement for each trait. This information will enable us to further examine the shared biological mechanisms between pigmentation phenotype and melanoma.