Sabanci University Machine Learning Project conducted under the supervision of Öznur Taştan.
Project scope: Deep mutational scanning (DMS) is a technique used by researchers to assess the effects of mutations on protein function. This method achieves to identify connections between mutations and their effects on molecular function and evolutionary processes. In this project, we focused on pinpointing specific mutations and their locations on the GB1 protein that would affect its interaction with IgG-Fc proteins. By implementing the MuMi (mutation and minimization) method, we systematically generated all possible single mutations of GB1. This comprehensive approach allowed us to construct a network graph, which then facilitated the development of an XGBoost model, which delivers superior outcomes compared to prior investigations in the field.
Developed with the collaboration of Faraz Badali, Büşra Tayhan, Alize Sevgi Yalçınkaya, and Bilgehan Cagiltay.
The data for this project has been sourced from: "T.F. Guclu, A.R. Atilgan, C. Atilgan, "Deciphering GB1's Single Mutational Landscape: Insights from MuMi Analysis," Journal of Physical Chemistry B; 128, 7987-7996 (2024)"