Language-guided domain adaptation improves protein sequence rep-resentations for antimicrobial resistance analysis
- 1Beijing Institute of Genomics, Chinese Academy of Sciences
Abstract
Antimicrobial resistance (AMR), particularly antibiotic resistance in bacterial pathogens, complicates infectious disease management. Because bacterial resistance is often mediated by protein functions, including antibiotic inactivation and efflux, protein sequence analysis provides a molecular basis for digital AMR surveillance and therapeutic intervention. We developed BacLLM, a language guided protein representation framework that integrates annotation-derived functional knowledge into protein language model representations for AMR analysis in WHO priority bacterial pathogens. Through sequence enhancement, semantic enrichment of UniProt annotations, and cross modal contrastive learning, BacLLM transfers LLM encoded annotation priors into the sequence encoder. Compared with baselines, BacLLM provided more informative representations for resistance-gene identification and mechanism prediction, captured resistance associated mutational signals at single-residue resolution, and preserved taxonomic relationships not reflected by phenotype based clinical groupings. Moreover, BacLLM recovered resistance associated functional organization within a TolC-centered protein association network, prioritizing proteins with low sequence identity to TolC but predicted structural similarity to TolC. Extending from AMR characterization to therapeutic prioritization, BacLLM also improved strain-level phage host matching for phage therapy candidate prioritization. Together, these findings establish BacLLM as a protein representation framework supporting AMR surveillance, functional characterization, and therapeutic candidate prioritization.
Related
- Phage-Inducible Chromosomal Jumbo Islands (PICJIs) spreading antimicrobial resistance, host-adaptation factors and defense systems, a novel satellite family of nucleus-forming Chimalliviridae
- Phage-ICE tandems facilitate bacterial adaptation and antimicrobial resistance spread
- Parameter-Efficient Adaptation of Protein Language Models for Dual-Target Antimicrobial Peptide Discovery Under Strict Homology Constraints
- π-MNovo improves de novo peptide sequencing through microbial-domain adaptation and evidence-guided candidate selection
- Adaptation in somatosensory afferents improves rate and temporal coding of vibrotactile stimulus features
BibTeX
@misc{ye2026languageguided,
title = {{Language-guided domain adaptation improves protein sequence rep-resentations for antimicrobial resistance analysis}},
author = {R. Ye and L. Li and Z. Geng and S. Song},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.04.764426},
url = {https://www.biorxiv.org/content/10.64898/2026.11.04.764426v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Language-guided domain adaptation improves protein sequence rep-resentations for antimicrobial resistance analysis
AU - R. Ye
AU - L. Li
AU - Z. Geng
AU - S. Song
PY - 2026
DA - 2026/11/04
PB - bioRxiv
DO - 10.64898/2026.11.04.764426
UR - https://www.biorxiv.org/content/10.64898/2026.11.04.764426v1
AB - Antimicrobial resistance (AMR), particularly antibiotic resistance in bacterial pathogens, complicates infectious disease management. Because bacterial resistance is often mediated by protein functions, including antibiotic inactivation and efflux, protein sequence analysis provides a molecular basis for digital AMR surveillance and therapeutic intervention. We developed BacLLM, a language guided protein representation framework that integrates annotation-derived functional knowledge into protein language model representations for AMR analysis in WHO priority bacterial pathogens. Through sequence enhancement, semantic enrichment of UniProt annotations, and cross modal contrastive learning, BacLLM transfers LLM encoded annotation priors into the sequence encoder. Compared with baselines, BacLLM provided more informative representations for resistance-gene identification and mechanism prediction, captured resistance associated mutational signals at single-residue resolution, and preserved taxonomic relationships not reflected by phenotype based clinical groupings. Moreover, BacLLM recovered resistance associated functional organization within a TolC-centered protein association network, prioritizing proteins with low sequence identity to TolC but predicted structural similarity to TolC. Extending from AMR characterization to therapeutic prioritization, BacLLM also improved strain-level phage host matching for phage therapy candidate prioritization. Together, these findings establish BacLLM as a protein representation framework supporting AMR surveillance, functional characterization, and therapeutic candidate prioritization.
ER -