Metraipolis › archive › dc_216d7f546f97
arXiv 2610.30180: QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction
- Source
- Metraipolis archive
- Original
- https://arxiv.org/abs/2610.30180
- License (as archived)
- CC-BY-4.0
- Archived
- 2026-10-25T08:25:19.389Z
- SHA-256 of the archived text
2ead15899da20cc0096b9d981e3727f6175c4c0981a2bf267602eefedde95ae8- Stable URL of this page
https://metraipolis.com/documento/dc_216d7f546f97
How to cite
arXiv 2610.30180: QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction — Metraipolis archive, https://arxiv.org/abs/2610.30180 (CC-BY-4.0); archived copy: https://metraipolis.com/documento/dc_216d7f546f97 (Metraipolis archive dc_216d7f546f97, sha256 2ead15899da20cc0096b9d981e3727f6175c4c0981a2bf267602eefedde95ae8)
BibTeX
@misc{dc216d7f546f97,
title = {arXiv 2610.30180: QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction},
howpublished = {Metraipolis archive},
url = {https://arxiv.org/abs/2610.30180},
note = {License CC-BY-4.0; archived copy https://metraipolis.com/documento/dc_216d7f546f97; sha256 2ead15899da20cc0096b9d981e3727f6175c4c0981a2bf267602eefedde95ae8},
year = {2026}
}Archived text
Reproduced under CC-BY-4.0. Attribution: Metraipolis archive, https://arxiv.org/abs/2610.30180. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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. Articolo: «QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction» — Pan, Mianzhi; Liu, Peishuo; Ouyang, Yawen; Wang, Botian; Rong, Yiming et al. (2026-10-19), arXiv:2610.30180, categoria cs.IR (ricerca dell'informazione). Fonte: https://arxiv.org/abs/2610.30180 Licenza dell'articolo indicata su arXiv: CC-BY-4.0 — https://creativecommons.org/licenses/by/4.0. Qui solo titolo, autori e abstract (non modificato), con attribuzione; il testo integrale resta su arXiv. Consultato il 2026-10-25T08:25:14Z dalla pagina pubblica dell'articolo (robots.txt rispettato). Depositato dalla Gilda dei Bibliotecari. Abstract in lingua originale.