33rd Annual Student Research Conference
Computational Detection of Synthetic Lethality for Cancer Treatment
Shreeya Rupakheti* and Deepson Shrestha
Dr. Jon Beck, Faculty Mentor
Abstract: A synthetic lethal interaction between two genes is when the perturbation of either gene alone remains viable, but the perturbation of both genes results in cell death. The detection of synthetic lethal pairs can be leveraged for cancer treatment. In this project we apply a Bayesian Group Factor Analysis (BGFA) model to conjointly analyze CRISPR-Cas9 and drug perturbations. The goal of this project is to computationally prioritize potential synthetic lethal gene pairs.
The BGFA model is a machine learning approach that learns a latent space representation where both input data types (CRISPR and drug response) are active. We will compile a “gold standard” resource from curated literature and existing databases to perform empirical validation. Promising synthetic lethal gene pairs will be submitted to cancer biologists at Oregon Health and Sciences University for experimental validation.
The computational platform is R code supplemented with Python for data preparation and results visualization.
Keywords: Synthetic Lethal , Cancer, Bayesian Group Factor Analysis, Drug response, CRISPR, Lethal Gene Pairs, R code, Machine Learning
Topic(s):Computer Science
Biology
Statistics
Presentation Type: Poster
Session: TBA
Location: TBA
Time: TBA