f-BGM

f-BGM enables accurate and interpretable genome-wide identification of biosynthetic gene clusters (BGCs) in fungal genomes. It is specifically tailored for fungal genomes, which often lack high-quality annotations. f-BGM combines multiple deep learning modules—including pre-trained domain-level (f-DLC) and sequence-level (f-ESM2) representations—to detect BGC boundaries and identify core biosynthetic enzymes. f-BGM integrates several open-source tools and pretrained models as part of its prediction workflow.

Zhou Y, Liu D, Huang Z, et al. f-BGM enables fungi-specific genome mining in high accuracy and interpretability. bioRxiv:2025.07.08.662937.