DeepSig is a web-server for predicting signal peptides and their cleavage sites. DeepSig is based on deep learning methods, in particular Deep Convolutional Neural Networks (DCNNs), specifically designed to accurately identify signal peptide sequences located at the N-terminus of the query protein sequence. Furthermore, DeepSig is optimized to discriminate between true signal peptides and similar N-terminal transmembrane segments. When a signal sequence is detected, the query sequence is analyzed by a probabilisic sequence labelling method (based on Conditional Random Fields) to identify the precise location of the cleavage site.

For more information see:

Savojardo C., Martelli P.L., Fariselli P., Casadio R. DeepSig: deep learning improves signal peptide detection in proteins, Bioinformatics (2018) 34(10): 1690-1696.

To start using DeepSig you simply need to paste protein sequences in FASTA format in the field below. You also need to specificy the kingdom the sequences belong to, chosing among: Eukaryotes, Gram-positive bacteria and Gram-negative bacteria. This because DeepSig has been optimized separately on each kingdom. Therefore, press the "Submit job" button to submit your job. The server accepts in input up to 500 protein sequences per submission. If you need to perform large-scale prediction, please, use the standalone version of DeepSig