🤖 AI Summary
Traditional molecular property prediction methods typically rely on a single conformation, overlooking the fact that molecules exist as ensembles of conformers in solution. This work proposes EnsembleEGNN, a novel model that represents the entire conformational ensemble as a unified embedding by encoding individual conformers with a shared equivariant graph neural network (EGNN) and aggregating ensemble information via a set attention mechanism. The model is pretrained through a multi-task self-supervised strategy involving masked token recovery, noisy coordinate reconstruction, and pairwise distance reconstruction, and is jointly optimized with a BERT sequence encoder. Evaluated on the CREMP-CycPeptMPDB dataset, EnsembleEGNN achieves an R² of 0.538 (Pearson r = 0.737), substantially outperforming existing baselines and demonstrating the efficacy and superiority of explicit conformational ensemble modeling for cyclic peptide property prediction.
📝 Abstract
Molecular property prediction from structure often uses a single representative conformation, even though many molecules exist as conformational ensembles in solution. We introduce EnsembleEGNN, a molecular ensemble foundation model that encodes an ensemble by first encoding each conformer with shared Equivariant Graph Neural Network (EGNN) layers, then pooling the resulting conformer representations with a Set Attention Block. We pretrain the model on CREMP, a cyclic peptide ensemble dataset, using a multi-task self-supervised objective combining masked token recovery, noisy-coordinate reconstruction, and pairwise distance reconstruction. On the CREMP-CycPeptMPDB dataset, training EnsembleEGNN from scratch fails entirely ($R^2=0.005$). However, the pretrained model reaches $R^2=0.477$ and Pearson $r=0.699$, outperforming the sequence-only BERT baseline ($R^2=0.439$, Pearson $r=0.667$). When EnsembleEGNN is co-trained end-to-end with the BERT sequence encoder, the hybrid model improves further to $R^2=0.538$ and Pearson $r=0.737$. These results demonstrate that encoding conformational ensembles into a single thermodynamically informed embedding improves cyclic-peptide property prediction.