How to increase sequence limit in SignalP

[JackMetal]

I'm using SignalP in my local computer. One thing that I've encountered is 'Sequence limit reached: Max 10000 sequences are allowed'. I've read through the instructions, and did not find the command to increase the sequence limit.
Is there a way to increase the sequence limit? Thank you very much!

[maubp]

Which version of SignalP are you using?

[LeightonP]

One option is to wrap SignalP (and companion programs) in a script that splits files for input, and concatenates output from several runs, e.g.

run_signalp.py

run_tmhmm.py

These two scripts also take advantage of multiple cores, where possible.

[maubp]

Some of this family of tools have the sequence limit as a setting inside the Perl wrapper script (which is why I was checking which version you are using).

But as Leighton points out, the more practical route is to take advantage of the embarrassingly parallel nature of the task and split the FASTA file and run multiple copies in parallel (on the same machine or a cluster).

We used the same trick in our SignalP and TMHMM wrappers for Galaxy http://toolshed.g2.bx.psu.edu/view/p...mm_and_signalp which we talk about a little in the accompanying paper: http://dx.doi.org/10.7717/peerj.167

[JackMetal]

Some of this family of tools have the sequence limit as a setting inside the Perl wrapper script (which is why I was checking which version you are using).

But as Leighton points out, the more practical route is to take advantage of the embarrassingly parallel nature of the task and split the FASTA file and run multiple copies in parallel (on the same machine or a cluster).

We used the same trick in our SignalP and TMHMM wrappers for Galaxy http://toolshed.g2.bx.psu.edu/view/p...mm_and_signalp which we talk about a little in the accompanying paper: http://dx.doi.org/10.7717/peerj.167

The version I'm using is signalp-4.1. I'm reading your paper. It looks awesome. I might have some further questions about the Galaxy.

[JackMetal]

One option is to wrap SignalP (and companion programs) in a script that splits files for input, and concatenates output from several runs, e.g.

run_signalp.py

run_tmhmm.py

These two scripts also take advantage of multiple cores, where possible.

Thank you so much. Can you give some direction on how to wrap signalp with the script you provided?

[LeightonP]

Thank you so much. Can you give some direction on how to wrap signalp with the script you provided?

It's pretty straightforward

1) Make sure SignalP is installed and working on your system.
2) Run the script.

run_signalp.py [-o|--outfilename <output file>] <euk|gram+|gram-><FASTAfile>

If you didn't make the script executable, then use "python run_signalp.py etc."

[JackMetal]

It's pretty straightforward

1) Make sure SignalP is installed and working on your system.
2) Run the script.

run_signalp.py [-o|--outfilename <output file>] <euk|gram+|gram-><FASTAfile>

If you didn't make the script executable, then use "python run_signalp.py etc."

Thanks. But when I run the script, it seems that I dont have all these modules, Bio, multiprocessing, os, sys etc. Does it mean I need to install all these modules listed to run the script.

[maubp]

Thanks. But when I run the script, it seems that I dont have all these modules, Bio, multiprocessing, os, sys etc. Does it mean I need to install all these modules listed to run the script.

Bio is the Biopython library, which you would need to install.

However, os, and sys are core Python modules in the standard library - and if they are missing your Python is very very broken. The multiprocessing library is also a core library, but only included in Python 2.6 onwards.

[j.provaz]

You can also edit the signalP code to allow you more sequences. If your computer can chew on it. Just edit the signalP file (around line 21) to some number of your choice:

# max number of sequences per run (any number can be handled)
my $MAX_ALLOWED_ENTRIES=1000;

[sindrle]

How can you do the same for SecretomeP ?

The reason I ask is because SecretomeP also runs SignalP. Very practical.

[Birdman]

j.provaz is right. Just editing the line "my $MAX_ALLOWED_ENTRIES=2000000;" does the trick.