Dannie Durand

The paper introduces a unified smoothing framework for protein domain bigram models, distinguishing between biologically incompatible and undersampled domain pairs.

This work analyzes CAFA5 submissions, showing improved accuracy in protein function prediction and highlights the role of protein language models and 3D structures in performance gains.

The paper reconstructs the protein interactome of the Last Eukaryotic Common Ancestor and identifies new human gene-disease associations for bone density and congenital birth defects.

The paper uses vector embeddings to model multidomain protein architectures and finds that contextual relationships between domains influence functional similarity.

The paper uses a Birth-Death-Gain model to analyze the evolution of protein domains in metazoa, finding that domains involved in metazoan innovations evolve faster and that domain repertoires undergo ongoing remodeling.