Summer Internship

My summer internship was with UCLA’s Bruins in Genomics Summer Program where I worked on determining the viability of using SHAP(Shapley Value Explanations) in determining how genotypes lead to phenotypes in complex machine learning models. The results we obtained from our research indicates that regardless of the complexity of our models and varying biological factors SHAP was able to identify casual loci for Linear Regression, Random Forest, and Neural Network models with high accuracy for one loci. When there were two loci Random Forest and Neural Nets both struggled to get both causal loci correct. We also found that Shapley values alone were not able detect interactions occurring between loci, but SHAP interaction values could be used to determine if interactions took place. From our results we can conclude that SHAP offers a promising method to identify casual loci and can be used to determine if interactions occurred between loci.

To view my summer report please use the following link Summer Report

To see my presentation for the summer program please use the following link Summer Presentation

To view the code created for the summer internship use the following link Summer Code