SignalP - choosing the right organism

Question

[6schulte]

Hi,

Has any of you experience with SignalP and can tell me whether "gram-" finds both gram negative and gram positive peptides?

One can run SignalP remote on their own PC in the command line like this:

signalp -fasta test.fasta -org gram-

With -fasta being a string representing the input file in fasta format. and -org being a string representing the input type of organism. Archaea: 'arch', Gram-positive: 'gram+', Gram-negative: 'gram-' or Eukarya: 'euk' (default "euk")

I want to catch any bacterial signal peptides. So I could just run it twice but would prefer not to ofc. :)

[6schulte]

If anybody comes across this and wants to know more about the input options, here is the link to SignalPs readme: http://www.cbs.dtu.dk/services/doc/signalp-5.0.readme

[Joe]

Judging by the documentation, it's only going to pick up gram negative if you provide the gram- flag. I don't know if its a required argument, maybe you can run without specifying, but I expect you're going to need to run it twice.

[6schulte]

Yes, I fears so too. My supervisor suggested otherwise but the documentation doesn't give any indications for this assumption so I will run a little test and see which of the two options finds the most signal peptides. If no significant difference between the two is found I will have to run it twice. Anyhow I will keep you updated on the results so that at least someone who comes across this in the future can benefit from this post :) Thanks for your reply!

Answer 1

[6schulte]

If anybody comes across this in the future:

I have compared the successrate of the three -org modes interesting for me: Archaea: 'arch', Gram-positive: 'gram+', Gram-negative: 'gram-'.

The result showed that there is 'no significant difference' (0) in success of finding/ identifying a signal peptide.

Nontheless some things that I have noticed:

• 'gram-' had the least signal pept. identified as “OTHER”

• 'gram+' was always part of the “majority” (1)

• 'arch' suggested the TAT signal pept. more often than the others, but less often the LIPO

I, for once, decided to run all three simultaneously (multi-threadding them) as SignalP processes the sequences very quickly anyhow....

================================================================

(0): I only used 119 sequences, all relatively similar, but from different organisms and of different length. This little survey is nothing fancy but helped me decide. Maybe it will hepl someone else too :)

(1): "part of the 'majority' ": meaning either agreed with 'gram-' or 'arch', when the pedictions implied different signal peptides. It was never the case that 'gram-' and 'arch' agreed on a signal pept. meanwhile 'gram+' sugested a different one.