- Secretory pathway (presence of a signal peptide SP).
- Mitochondrion (presence of a mitochondrial targeting peptide mTP)
- Chloroplast (presence of a chloroplast transit peptide cTP; only applicable to plant proteins)
- Other (nucleus, cytoplasmic, or otherwise).
Paste or upload your sequences below. Indicate if proteins are plant or non-plant. Press submit. If you provide 20 or less sequences you will be presented with both an overall prediction of the sequence target and a target detection score on the residue level. If more than 20 sequences are provided only the overall sequence score is presented.
The subcellular localisation predictor is largely based on TargetP. Version 1.0 is trained and tested on the same training data which can be retrieved from the TargetP web site. The predictor is presented and evaluated in the following paper.
Bodén, M. and Hawkins, J. (Submitted). Prediction of subcellular localisation using sequence-biased recurrent networks.
The target detection score on the residue level is presented in
Bodén, M. and Hawkins, J. (In press). Detecting residues in targeting peptides. In Proceedings of the Asia-Pacific Bioinformatics Conference, Singapore.
A comprehensive study of recurrent networks for sequence analysis and for subcellular localisation in particular is presented in
Hawkins, J. and Bodén, M. (Accepted). An architectural bias for accessing patterns in biological sequences. IEEE/ACM Transactions on Computational Biology and Bioinformatics.
This service has been developed by Mikael Bodén, John Hawkins and James Watson.