Decoding DNA methylation across human tissues
Melody predicts locus-specific DNA methylation from genomic sequence, capturing both local and long-range signals across 39 human tissues. Its extension, Melody-G, integrates transcriptomic information to infer methylation in previously unseen cell types.
Automated, interpretable biological sequence analysis
A platform for sequence-level prediction, base-level functional annotation and visual interpretation. DeepBIO brings together 42 deep learning methods to help researchers train, compare and interpret models for DNA, RNA and protein sequences.
Interpretable prediction of DNA methylation
A multi-scale biological language learning model for predicting DNA methylation directly from genomic sequences. iDNA-ABF combines language representations with an attention-based feature fusion network to identify methylation sites and explore their sequence context.
Explainable learning of peptide secondary structure
PHAT combines a pretrained protein language model with hypergraph learning to predict three-state and eight-state peptide secondary structures. Its interpretable predictions support further exploration of peptide structure and function.
Structure-aware protein binding site prediction
DeepProSite combines ESMFold structures, pretrained protein language representations and a graph transformer to identify binding residues. It supports the prediction of sites involved in protein, peptide, nucleic acid and other ligand interactions.
From peptide detection to structure and function
An integrated platform for peptide detectability, secondary structure and functional prediction. DeepPEP combines deep learning models with visual analysis to help explore bioactive peptides and their potential therapeutic properties.
Human-guided, AI-assisted retrosynthesis
An interactive platform for retrosynthetic analysis that brings AI-assisted synthesis planning together with human expertise. Explore routes from a target molecule and refine the planning process through an interactive workflow.
Questions about our tools? Get in touch
WeiLab 2021 by WEIlab