Protein Thoughts: Interpretable Reasoning with Tree of Thoughts and Embedding-Space Flow Matching for Protein-Protein Interaction Discovery

TL;DR

Protein Thoughts introduces an interpretable, hypothesis-guided search framework for protein-protein interaction discovery, combining explicit reasoning, embedding-space flow matching, and a language model-guided search strategy.

Contribution

The paper presents a novel, interpretable framework that decomposes PPI evidence into meaningful signals and employs a hypothesis-guided Tree-of-Thoughts search with embedding-space flow matching.

Findings

Achieves a mean best-binder rank of 11.2 on SHS148k benchmark, outperforming baseline methods.

Attains 91.08% Micro-F1 in binding prediction, surpassing existing PPI approaches.

Demonstrates 76% improvement over entropic tree search baseline in binder ranking.

Abstract

Protein-protein interactions (PPIs) govern nearly all cellular processes, yet computational methods for identifying binding partners typically produce ranked predictions without mechanistic justification. This creates a fundamental barrier to adoption because biologists cannot assess whether predictions reflect genuine biochemical insight or spurious correlations. We present \textbf{Protein Thoughts}, a framework that reformulates PPI discovery as an interpretable search problem with explicit reasoning. The system decomposes binding evidence into four biologically meaningful signals: sequence similarity reflecting evolutionary relationships, structural complementarity capturing geometric fit, interface balance, and chemical compatibility encoding residue-level interactions. Rather than collapsing these signals into an opaque score, we preserve their individual contributions through a transparent value function that enables both ranking and auditing. To navigate large candidate spaces efficiently, we introduce hypothesis-guided entropy-regularized Tree-of-Thoughts search. A fine-tuned language model generates search directives from embedding-derived features, classifying candidates as high-priority, exploratory, or skippable. These directives condition a Boltzmann policy that balances exploitation with entropy-driven exploration, while hypothesis-aware pruning prevents premature abandonment of promising candidates. For candidates exhibiting score disagreement, hypothesis-conditioned embedding-space flow matching transports protein embeddings toward the binder manifold. On the SHS148k benchmark, Protein Thoughts achieves mean best-binder rank of 11.2 versus 47.7 for an entropic tree search baseline, a 76% improvement, and for binding prediction the trained value function achieves $91.08 \pm 0.19$ Micro-F1, outperforming existing PPI methods on the same dataset.