Скачать или смотреть How to perform fast and accurate 3D ligand-based affinity predictions

Binding affinity prediction continues to be a challenge in computer-aided drug design, especially in the absence of a high-quality target structure. To address this limitation, we present QuanSA, a 3D QSAR-based method that leverages machine learning algorithms to accurately estimate binding affinities in the absence of protein structural data and with limited ligand activity information.

Join Himani Tandon, PhD, to explore the theory and methodology behind QuanSA, see case study examples that demonstrate its application in lead optimisation scenarios, and learn how you can:

Develop physically realistic and causal models for affinity prediction
Extrapolate outside the chemical space of your training set
Iteratively refine the accuracy and domain coverage of existing models
Build informative models by leveraging the effective conformational search and accurate multiple structure alignment
Incorporate confidence metrics to identify novel active compounds with greater reliability
Enhance the performance of free energy perturbation (FEP+) calculations using an integrated approach, to improve resource management and decision making.

Chapters:
0:00 Introduction to the speaker
0:54 Overview
2:14 Introduction to QuanSA
25:00 QuanSA benchmarking
40:34 QuanSA applications
51:16 Q&A