Predicting variant-induced drug resistance by combining artificial intelligence with deep mutational scanning and 3D protein-drug complex structures
| Title: |
Predicting variant-induced drug resistance by combining artificial intelligence with deep mutational scanning and 3D protein-drug complex structures |
| DNr: |
NAISS 2024/22-752 |
| Project Type: |
NAISS Small Compute |
| Principal Investigator: |
Yoomi Park <yoomi.park@ki.se> |
| Affiliation: |
Karolinska Institutet |
| Duration: |
2024-06-05 – 2025-06-01 |
| Classification: |
10201 |
| Keywords: |
|
Abstract
Genetic variants can influence drug efficacy by altering the binding affinity of drugs to their protein targets. To understand how specific genetic variants disrupt protein-drug binding and impact drug efficacy, our project integrates artificial intelligence with deep mutational scanning and structural analysis of 3D protein-drug complexes. The project has four key objectives: (a) experimentally determining the functionality of all possible MDR1 transporter variants across multiple drugs, (b) calculating the impact of genetic variants on drug binding by linking 3D protein structures with genetic data, (c) developing AI models to predict drug efficacy changes due to genetic variants, including those not experimentally investigated, and (d) implementing a clinical decision support system for precision dosing and drug selection. This approach empowers physicians to formulate optimal management strategies, advancing personalized medicine and enhancing healthcare competitiveness.
