Kexin Xu

Kexin Xu (Oxford, England) is a Ph.D. student in Computational Biology at the University of Oxford working on developing deep learning models to predict (single-cell) spatial transcriptomics (gene expression map) from gut histopathology images, with research experience in medical imaging, epidemiology, multi-omics, and natural language processing.

She holds a B.Sc. in Medical Biosciences from Imperial College London. She ranked first for two consecutive years in her undergraduate degree and can code in Python, MATLAB, and R. The Equitech Futures Institute consolidated her determination to pursue AI for equitable healthcare.

Supervised by Dr. Abhilash Mishra, she worked previously with fellow Equitech Scholars Aryan Gupta, Eugene Herasimov, Yash Gorana, and Kartik Chincholikar on the Nightingale Competition to predict breast cancer from biopsy images.

Some of her previous projects include:

1) developing a novel random forest model for correlated health datasets for enhanced feature importance (the model got presented at the European Computational Biology Conference 2022);
2) developing a pipeline for the in-silico mutagenesis and ligand binding analysis of uncrystallised proteins using AlphaFold2 (the project won the Best Industrial Prize in the International Directed Evolution Competition 2021);
3) developing an ageing clock using epigenomics & metabolomics to predict biological age.