SEQ2FUN

[an error occurred while processing this directive]


StarFunc is a composite method combing structure, sequence, protein-protein interaction (PPI), Pfam family, and deep learning-based approaches for robust biological function annotation of proteins. Starting from the 3D structural model, StarFunc will thread the query structure through the AlphaFold DB and BioLiP protein function database. Meanwhile, sequence homologs, PPI partners and protein families will be detected from the UniProt-GOA, STRING and Pfam databases. Template-free function annotations are also performed using a deep learning model. Those five independent predictions are combined using random forest models to generate consensus function predictions in the form of Gene Ontology (GO) terms. StarFunc (as team hfm7zc) was ranked among the top teams in the recent CAFA5 challenge. Tutorial for the usage of StarFunc can be found at the help page. Questions about the StarFunc server can be posted at the Service System Discussion Board. (Prediction for the Human reference proteome) (Example output) (Help page)

  • Structue of query protein ([30,1500] residues in PDB format. Copy and paste your structure file here    Sample structure input

      Or upload the stucture file from your local computer
  • Email: (Though optional, an email address is highly recommended to ensure prompt notification of job status)
  • Job ID: (Optional, your given name for this protein)

    • References:
    • Chengxin Zhang, Quancheng Liu, Lydia Freddolino (2024) StarFunc: fusing template-based and deep learning approaches for accurate protein function prediction. bioRxiv [preprint]