QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction
En palabras de los autores
Predicting protein functions from amino acid sequences is essential to understanding biological processes. Large language models (LLMs) offer flexible function descriptions to natural-language queries, yet their limited protein understanding motivates retrieval-augmented generation (RAG). Existing protein RAG methods primarily expand external retrieval signals while overlooking the rich knowledge encoded within a single protein language model (PLM). We show that retrieval performance varies across PLM layers and biological tasks, making conventional final-layer selection suboptimal. To this end, we propose QProtFuse, which uses queries to adaptively fuse multi-layer protein features and jointly learns protein and annotation retrieval through intra-modal and cross-modal contrastive objectives. On Prot-Inst-OOD, QProtFuse achieves the best protein and annotation retrieval scores, reaching 16.4% Recall@1 with annotation reranking. These gains improve downstream RAG-based function descriptions, yielding the highest LLM-as-a-judge scores for both Qwen3-8B (6.22/10) and DeepSeek-V4-Pro (6.68/10). Ablations and weight analyses demonstrate the value of query-guided fusion and complementary training objectives for exploiting a PLM's internal knowledge.
Apareció: viernes, 23 de octubre. arXiv. Preprint, todavía sin revisión por pares.