July 14, 2025Open Access

EZpred: improving deep learning-based enzyme function prediction using unlabeled sequence homologs

View Full Paper

Ask AI

Bookmark

EZpred is the first deep learning model to effectively use unlabeled sequence homologs for protein function prediction.
It achieved a 4% higher F1-score for EC number prediction compared to similar models not using sequence homologs.
Features were extracted using the ESMC protein language model, optimizing input from homologous sequences.
This research underscores the potential of overlooked sequence homolog features in advancing protein function prediction.

Discussion

No takes yet. Share an insight, caveat, or question.

Deep learning model improves enzyme function prediction accuracy using sequence homologs, highlighting their crucial role in feature extraction.

Zhang et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Deep learning model improves enzyme function prediction accuracy using sequence homologs, highlighting their crucial role in feature extraction.