arXiv:2609.00518v1 Announce Type: new Abstract: Antibody-specific language models pretrained via masked language modeling (MLM) learn representations that are critical for downstream sequence design...
ProtLingo is a protein language modeling framework that enhances a pretrained single‑sequence Transformer backbone with conditional local memory and sparse expert routing. It maps residue representations into discrete codes, composes local windows into latent N‑gram addresses, and retrieves reusable residual signals for recurring sequence contexts. The model also converts selected feed‑forward blocks into sparse Mixture‑of‑Experts layers, allowing residue‑dependent computation while activating only a subset of parameters, achieving competitive performance on protein fitness prediction, FLIP benchmarks, and supervised contact prediction with a 150M‑parameter backbone.
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:2607. 22777v1 Announce Type: cross Abstract: Protein language models learn transferable sequence representations.
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.
SymFold introduces a symmetric dual‑path architecture that combines protein language models (PLMs) and multimodal protein language models (MPLMs) to iteratively guide protein sequence generation for inverse folding. By leveraging pretrained sequence evolution knowledge from PLMs and structural knowledge from MPLMs, the method improves upon the traditional serial pipeline where structure encoders produce coarse sequences refined by PLMs. Experiments on standard inverse‑folding benchmarks show state‑of‑the‑art performance, and ablation studies confirm the effectiveness of the symmetric design.
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.
arXiv:2602. 23179v4 Announce Type: replace Abstract: Protein sequences are abundant in repeating segments, both as exact copies and as approximate segments with mutations.
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.
arXiv:2603.29529v2 Announce Type: replace-cross Abstract: Using a statistical mechanics framework, we investigate the parameter space of transformer models trained on protein sequence data. We sample...
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.
arXiv:2608. 12090v1 Announce Type: new Abstract: Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology.