Coarse composition suffices: tabular in-context learning for multi-activity antimicrobial peptide profiling
- 1IIT Roorkee
Abstract
Antimicrobial peptides (AMPs) often act against multiple pathogen classes, making multi-label activity prediction a more realistic screening target than binary antimicrobial classification. The ESCAPE benchmark formalizes this setting, but leading approaches typically rely on multimodal, structure-conditioned deep models that are costly to train and tune. We show that a simple, sequence-only pipeline can match and surpass these methods by combining 330 interpretable sequence descriptors with TabPFN, a tabular foundation model that performs in-context prediction in a single forward pass without gradient-based training or hyperparameter search. On ESCAPE (82,359 peptides; five labels), a label-powerset TabPFN model achieves mAP-5=77.8%, improving on the previously best reported 72.1%. A probabilistic classifier chain is the first method to match or exceed the best published average precision on each of the five labels simultaneously. The gains persist under the prior state-of-the-art single-fold training protocol, indicating they are not a training-set-size artefact, and are largest for remote homologues (+11.2 points below 30% sequence identity). Ablations further show that predicted structure is unnecessary at inference and that performance is not driven by any single descriptor family: ten global physicochemical scalars recover 91% of full-feature performance. Finally, explicitly modelling label dependence yields targeted benefits for scarce activities and supports ranking which activity to assay next from partial positive evidence.
Related
- Developmental Dynamics of Bioactive Composition and Biological Activities of Methanolic Extracts from Orange (Citrus sinensis) Peel: Comprehensive Metabolite Profiling, Antioxidant Capacity, and Antibacterial Activity
- CIT8, a machine learning-optimized antimicrobial peptide, exhibits potent activity against carbapenem-resistant Acinetobacter baumannii
- Integrated melissopalynological and phenolic profiling reveals antioxidant and antimicrobial potential of Turkish honeys
- Chemical and Enantiomeric Profiling of Melaleuca viridiflora Sol. ex Gaertn. Essential Oil and Evaluation of Its Antimicrobial, Antioxidant, and Anticholinesterase Activities
- Bacterial Periplasmic Oxidoreductases Control the Activity of Oxidized Human Antimicrobial β-Defensin 1
BibTeX
@misc{pal2026coarse,
title = {{Coarse composition suffices: tabular in-context learning for multi-activity antimicrobial peptide profiling}},
author = {A. Pal and R. Kumar and D. Solanki and P. Pareek and J. Singh and J. Singla},
howpublished = {bioRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.27.747591},
url = {https://www.biorxiv.org/content/10.64898/2026.08.27.747591v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - Coarse composition suffices: tabular in-context learning for multi-activity antimicrobial peptide profiling
AU - A. Pal
AU - R. Kumar
AU - D. Solanki
AU - P. Pareek
AU - J. Singh
AU - J. Singla
PY - 2026
DA - 2026/08/27
PB - bioRxiv
DO - 10.64898/2026.08.27.747591
UR - https://www.biorxiv.org/content/10.64898/2026.08.27.747591v2
AB - Antimicrobial peptides (AMPs) often act against multiple pathogen classes, making multi-label activity prediction a more realistic screening target than binary antimicrobial classification. The ESCAPE benchmark formalizes this setting, but leading approaches typically rely on multimodal, structure-conditioned deep models that are costly to train and tune. We show that a simple, sequence-only pipeline can match and surpass these methods by combining 330 interpretable sequence descriptors with TabPFN, a tabular foundation model that performs in-context prediction in a single forward pass without gradient-based training or hyperparameter search. On ESCAPE (82,359 peptides; five labels), a label-powerset TabPFN model achieves mAP-5=77.8%, improving on the previously best reported 72.1%. A probabilistic classifier chain is the first method to match or exceed the best published average precision on each of the five labels simultaneously. The gains persist under the prior state-of-the-art single-fold training protocol, indicating they are not a training-set-size artefact, and are largest for remote homologues (+11.2 points below 30% sequence identity). Ablations further show that predicted structure is unnecessary at inference and that performance is not driven by any single descriptor family: ten global physicochemical scalars recover 91% of full-feature performance. Finally, explicitly modelling label dependence yields targeted benefits for scarce activities and supports ranking which activity to assay next from partial positive evidence.
ER -