The paper introduces a method that uses orthogonal projection to remove the influence of known biochemical features from protein language model (PLM) embeddings, allowing the authors to assess how much these features contribute to protein fitness predictions. By applying this technique to high‑order and interaction effects, they demonstrate that eliminating these interpretable features reduces downstream classifier performance, indicating that PLM embeddings encode patterns correlated with biochemical properties. The authors also show that these biochemical features explain a substantial portion of the variance in the classifier’s predictions, suggesting that PLM embeddings capture biologically relevant information.
arXiv:2608. 12090v1 Announce Type: new Abstract: Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology.
Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology. These models, trained on large corpora of protein sequence data, are widely used to translate amino acid sequences into latent-space embeddings, ready for use in diverse downstream tasks (DTs).
The paper introduces a scalable method to interpret sparse autoencoder (SAE) features in the ESM-2 protein language model by leveraging geometrically inspired features of the protein α‑carbon backbone. Across 8M layers of ESM-2, a false discovery rate–controlled analysis shows that local geometry is significantly associated with many SAE features, revealing substructure within known biological labels and enabling annotation of unannotated metagenomic proteins. Ablation experiments demonstrate that removing these geometric features shifts ESM-2’s predicted contact maps toward the descriptor, linking mechanistic interpretability with structural biology.
arXiv:2606. 31126v1 Announce Type: new Abstract: Predicting biomolecular properties from limited labeled data is a central bottleneck in protein engineering and small-molecule design.
arXiv:2606. 16044v1 Announce Type: new Abstract: Protein language models (pLMs) can generate novel protein sequences with properties beyond those observed in nature, yet the mechanisms underlying protein generation remain poorly understood.
arXiv:2606. 02629v1 Announce Type: cross Abstract: Protein-protein interactions (PPIs) are essential for many biological processes.
arXiv:2606. 27440v1 Announce Type: new Abstract: Foundation models for structural biology have achieved remarkable performance in predicting biomolecular structure and show promise for the design of proteins and small molecules.
arXiv:2512. 15133v3 Announce Type: replace-cross Abstract: Proteins inherently possess a consistent sequence-structure duality.
arXiv:2606. 11057v1 Announce Type: new Abstract: Despite its importance to applications in protein design, predicting protein properties like binding affinity and thermostability from sparse experimental data remains a significant challenge.
arXiv:2605. 01625v3 Announce Type: replace Abstract: Proteins are inherently multiscale physical systems whose functional properties emerge from coordinated structural organization across multiple spatial resolutions, ranging from atomic interactions to global fold topology.
The paper investigates how guided protein language models can collapse onto off‑manifold representations when heavily steered to optimize a property. This collapse causes generated sequences to become low‑complexity and statistically similar to random amino‑acid input, yet the property oracle may still rate them highly. The authors propose a cheap, training‑free Mahalanobis filtering step that removes such off‑manifold candidates, improving both property scores and structural plausibility without altering the generator.
The paper investigates a problem in guided protein language models where strong guidance causes the model’s internal representations to collapse onto a region indistinguishable from random amino‑acid input, leading to low‑complexity sequences that still score well on the targeted property. The authors identify this off‑manifold collapse as a detectable signature and propose a post‑hoc filtering technique—Mahalanobis filtering—that removes atypical candidates based on a density prior over natural activations. This simple, training‑free step improves both property scores and structural plausibility across different guidance methods without altering the generator.