SPPIDER-seq is a tool for partner-specific prediction of protein-protein interaction (PPI) sites using only sequence information.
Given a query protein together with one or more candidate interaction partners (as amino acid sequences),
it uses pretrained models (based on ESM-2 embeddings) to predict which residues on the query protein are likely to participate in the interface with the partner(s).
The output is a per-residue score or label indicating putative interaction sites,
allowing users to see which parts of the protein are predicted to mediate binding with each specific partner.
Currently, SPPIDER-seq is not implemented as a web-server. Use a stand-alone version below.
Download the SPPIDER-seq Jupyter notebook from GitHub
Run the SPPIDER-seq Jupyter notebook on Google-Colab
