Vancouver, British Columbia, Canada
https://vvsection.com
MAT135, MAT136, MAT232, MAT236, STA107 Teaching Assistant Lecturing, conducting office hours, preparing class materials/tutorials/tests/exams, grading assignments/exams, communicating with TAs.
Study of Multi-Particle Diffusion Limited Aggregation (MDLA) - stochastic model of infection spread. This project compiled the tools needed to understand the model, such as the foundations of measure-theoretic probability, and culminated in a proof of a theorem concerning the growth of MDLA in a certain space. Supervised by Prof. Duncan Dauvergne.
• Explored various biophysical and genetic features of low-complexity regions in the human proteome • Established prediction biases for proteins with low-complexity regions in AlphaFold and ProstT5 (Foldseek) protein language models • Traced the differences that alternative splicing induces in predicted protein structures with regards to low-complexity regions • Identified that cancer mutations are significantly less likely to happen on low-complexity regions • Created data analysis pipelines for processing genomic information from gnomAD and The Cancer Genomic Atlas
Conceptualizing and testing dimensionality reduction methods for protein simulation data to increase interpretability. Performing quantitative and qualitative analyses of a dimensionality reduction method based on variational autoencoders. Building Markov state models of protein simulations to discover the underlying dynamics. Utilizing the University’s high-performance computing resources (Niagara and MIST) to train neural networks and run extensive protein simulations.