Maria Kadukova

The paper develops a method combining protein language models with 3D tesselation to create detailed binding site representations, improving downstream tasks like druggability modeling and benchmarking.

The paper introduces Convex-PLR, a physics-informed machine learning model for virtual screening that improves scoring function performance and interpretability.

The paper introduces VoroCNN, a deep convolutional neural network for predicting local qualities of 3D protein folds using Voronoi tessellation, achieving competitive results compared to existing methods.