Xin Dai

PEACE improves effector protein prediction by using contrastive embeddings and prototype-aware training, outperforming EffectorP 3.0 on fungal and fungi+oomycete datasets.

The paper develops ESMBind, a workflow combining evolutionary and physics-based modeling to predict metal-binding proteins and their structures, showing improved performance over existing methods.

This work integrates AlphaFold with a deep learning tool, CryoFEM, to improve the accuracy of 3-D atomic models built from cryo-EM maps.