Layer Probing Improves Kinase Functional Prediction with Protein Language Models

TL;DR

This study shows that analyzing all layers of protein language models, especially mid-to-late layers, enhances kinase function prediction accuracy compared to using only final-layer embeddings.

Contribution

The paper systematically evaluates all layers of ESM-2, revealing that intermediate layers contain valuable biological signals for kinase function prediction, improving upon previous methods.

Findings

Mid-to-late layers outperform final layer in unsupervised clustering.

Supervised accuracy improves to 75.7% with layer selection.

Reproducible benchmarking pipeline enhances reliability.

Abstract

Protein language models (PLMs) have transformed sequence-based protein analysis, yet most applications rely only on final-layer embeddings, which may overlook biologically meaningful information encoded in earlier layers. We systematically evaluate all 33 layers of ESM-2 for kinase functional prediction using both unsupervised clustering and supervised classification. We show that mid-to-late transformer layers (layers 20-33) outperform the final layer by 32 percent in unsupervised Adjusted Rand Index and improve homology-aware supervised accuracy to 75.7 percent. Domain-level extraction, calibrated probability estimates, and a reproducible benchmarking pipeline further strengthen reliability. Our results demonstrate that transformer depth contains functionally distinct biological signals and that principled layer selection significantly improves kinase function prediction.