BioBlobs: Differentiable Graph Partitioning for Protein Representation Learning

Xin Wang, Carlos Oliver

Keywords: protein representation learning, graph partitioning, graph neural networks

TL;DR: BioBlobs learns a shared codebook of flexible, motif-level partitions of proteins — improving encoder performance while providing interpretable, function-relevant substructures

Abstract:

Protein function is driven by coherent substructures which vary in size and topology, yet current protein representation learning models (PRL) distort these signals by relying on rigid substructures such as $k$-hop and fixed radius neighborhoods.
We introduce $\textrm{BioBlobs}$, a plug-and-play, fully differentiable module that represents proteins by dynamically partitioning structures into flexibly-sized, non-overlapping substructures (``blobs”).
The resulting blobs are quantized into a shared and interpretable codebook, yielding a discrete vocabulary of function-relevant protein substructures used to compute protein embeddings.
We show that $\textrm{BioBlobs}$ representations improve the performance of widely used protein encoders such as GVP-GNN across various PRL tasks.
Our approach highlights the value of architectures that directly capture function-relevant protein substructures, enabling both improved predictive performance and mechanistic insight into protein function.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 22839