AMAPEC is a machine learning framework developed in the Thomma lab to predict antimicrobial activity in fungal secreted effector proteins. The tool addresses a key limitation in effector biology: the lack of methods tailored to large, structurally diverse fungal effectors, which are poorly captured by classical antimicrobial peptide predictors.

Unlike peptide-focused approaches, AMAPEC is trained on a curated dataset of experimentally validated antimicrobial proteins, explicitly designed to accommodate the size and complexity of fungal effectors. It integrates sequence-derived features, physicochemical properties, and structure-informed descriptors to classify proteins based on their likelihood of antimicrobial activity.

AMAPEC thereby provides a scalable framework to explore antimicrobial effector repertoires and to generate hypotheses on how fungi shape their biotic environments through secreted proteins.

For a detailed description, see:
https://www.biorxiv.org/content/10.1101/2024.01.04.574150v1

Please cite: Mesny et al. (2026). Plant-associated fungi co-opt ancient antimicrobials for host manipulation. Sci Adv. 12: eaec1406. 

AMAPEC is made available through GitHub with a GPL v3.0 license: 
https://github.com/fantin-mesny/amapec.

Additionally, we provide a Google Colab notebook to try AMAPEC and to perform online antimicrobial activity prediction:
https://colab.research.google.com/github/fantin-mesny/amapec/blob/main/googleColab/AMAPEC.ipynb