Biology is complex. Proteins are the major players in biological processes. Proteins don’t work in isolation, they form complex interaction networks. Understanding their interactions and signaling are key in relating genetic and somatic variations to diseases. Our group is a pioneering group in developing methods for predicting structural models of protein-protein interactions to unravel genotype-phenotype relations. The unprecedented explosion of biological data and recent/impressive advances in machine learning provide opportunities to deal with complex biological problems at a large scale. The challenges are
1) how to integrate diverse and large biological data with deep learning to predict the structure of protein-protein interactions at atomic detail at genome-scale, and
2) how to infer the impact of mutations on interactions
3) more efficiently and with better accuracy. Join us to explore deep learning to attack this challenge.
The candidate will conduct research on
1) New deep learning models to represent protein space using sequence and 3D structural data. The 3D structural data is particularly challenging.
2) Explainable deep learning models for predicting protein-protein interactions
2) Deep learning models to integrate dynamics such as using data from molecular dynamics or coarse grain models.
The candidate is expected to have a strong background in machine learning and be familiar with the molecular biology of proteins/genes.
Koc University and KUIS AI center provides an excellent interdisciplinary environment involving engineering, science, and medical schools.