About this Abstract |
Meeting |
MS&T23: Materials Science & Technology
|
Symposium
|
Materials Processing and Fundamental Understanding Based on Machine Learning and Data Informatics
|
Presentation Title |
AI/ML Aided Drug Biomolecule and Materials Design |
Author(s) |
Mehdi Yazdani-Jahromi, Ali Khodabandeh Yalabadi, Aida Tayebi, Niloofar Yousefi, Elayaraja Kolanthai, Craig J. Neal, Sudipta Seal, Ozlem Ozmen Garibay |
On-Site Speaker (Planned) |
Mehdi Yazdani-Jahromi |
Abstract Scope |
In-silico virtual screening approaches can be time-consuming and expensive. Artificial intelligence has been widely applied across different stages of drug biomolecule design and development and has significantly reduced the time and cost involved in the process. In our previous study, we proposed BindingSite-AugmentedDTA. This framework is highly generalizable, interpretable, and can be combined with any deep learning-based regression model and can improve Drug-Target Affinity (DTA) predictions by finding the most probable binding sites of protein and narrowing down the search space. Our computational results show performance improvement in the seven state-of-the-art DTA models by up to 20%.
Furthermore, 13 compounds were used as experimental validation. AI/ML-designed biomolecules are added to nano-surfaces for pathogen mitigation. These measurements were in high agreement with computationally predicted results and supported the claim that Utilizing this framework as a tool has the potential to accelerate drug discovery and other biomaterials design processes. |