Seq-InSite
Sequence-based Interaction Site prediction
Welcome to the Seq-InSite web server for protein interaction site prediction.
Follow the steps below to use the server. We provide also the possibility of
running our previous program, PITHIA, by selecting it from the menu below.
Usage:
1. Choose the prediction model from the menu: Seq-InSite (default) or PITHIA.
2. Provide the e-mail address to receive the results.
3. Input the protein sequences.
- Maximum 10 sequences at the time,
each with length between 31 and 1000 amino acids.
- The sequences must be in FAST format, e.g.:
>WP_058509682.1 phosphoenolpyruvate carboxylase [Legionella steelei]
MNPNVSHLPRELSHMVSWSLTRLGKAIADVYGKETYDRIEQIRLSMQDTIGNKTFALRKA
...
>WP_085304639.1 phosphoenolpyruvate carboxylase [Colwellia polaris]
MLETISQNIRSLGSVLGKTIANDKGQRWLENVERVRVLAKDGVAQDALAINELQTLFESC
...
4. Click the "predict" button
Upon submitting your request, the server will show an estimate time required
to generate the results, and once the calculations are complete,
the results will be delivered to you via email.
Source Code and Data
The Seq-InSite source code, pre-trained models,
and all training and testing datasets are available on
GitHub
Citation
S. Hosseini, G.B. Golding, L. Ilie, Seq-InSite: sequence supersedes
structure for protein interaction site prediction, submitted, 2023.
S. Hosseini, L. Ilie, L. PITHIA: Protein Interaction Site Prediction
Using Multiple Sequence Alignments and Attention. Int. J. Mol. Sci.
2022, 23, 12814. https://doi.org/10.3390/ijms232112814
Acknowledgement
The backend server of Seq-InSite is provide by Compute Canada.