Accurate Predictions of Molecular Properties of Proteins via Graph Neural Networks and Transfer Learning.

ABSTRACT: Machine learning has emerged as a promising approach for predicting molecular properties of proteins, as it addresses limitations of experimental and traditional computational methods. Here, we introduce GSnet, a graph neural network (GNN) trained to predict physicochemical and geometric properties including solvation-free energies, diffusion constants, and hydrodynamic radii, based on three-dimensional protein structures. By leveraging transfer learning, pretrained GSnet embeddings were adapted to predict solvent-accessible surface area (SASA) and residue-specific pK_a values, achieving high accuracy and generalizability. Notably, GSnet outperformed existing protein embeddings for SASA prediction and a locally charge-aware variant, aLCnet, approached the accuracy o

SUBMITTER: Wozniak S 

PROVIDER: S-EPMC12080100 | biostudies-literature | 2025 May

REPOSITORIES: biostudies-literature