Over the summer of 2018, I worked on with Dr. Qingrun Zhang and the University of Calgary's bioinformatics department to evaluate the efficacy of existing cross-platform, cross-species and cross-tissues algorithms that evaluate genetic similarity. This project stores modified code from Tamayo et. al. regarding metagene projection, a cross-species method that utilized non-negative matrix factorization.
More information regarding HYRS can be found here: https://www.ucalgary.ca/utoday/issue/2018-08-10/budding-scientists-indulge-their-curiosity-during-summer-science-research-program
- R: used for data wrangling, matrix factoring, pseudoinverse projections, model construction and visualizations
- Linux: converted hundreds of .txt files to .GCT and .CLS files.
- Non-negative matrix factorization
- Pseudoinverses
- Hierarchical clustering
- Support vector machine