This is a package designed to generated secretome predictions by combining outputs of various programs (packaged in dependencies).

Currently it is hardcoded for eukaryotic fungal sequences.

Specifically the program:

• Uses signalp to identify proteins with signal peptides
• Checks the mature sequences of these proteins to ensure there are no TM domains (TM domains in the signal peptide itself will pass this stage) via tmhmm
• Identifies signal peptides targeted for secretion using targetp
• Identifies sequences likely to be extracellular compartment using wolfpsort

By default predictions are combined conservatively i.e. the predicted secretome is only those sequences that fulfill all of the above criteria but the pipe can optionally be run permissively where passing any stage will be sufficient for output in the predicted secretome e.g. has signal peptide for secretion but is not predicted as extracellular, or has a signal peptide but has TM domains in the mature sequence etc.

The package can also optionally output predicted transporters based on a minimum number of TM domains appearing in a sequence (or mature sequence in the case of proteins with signal peptides). This optional output and TM domains threshold can be specified with '-t NUMBER'

A full analysis can be run using the secretome_pipe.py script (all options can be shown using use -h)

  -h, --help                                  Show help message and exit
  --fasta INPUT_FILE, -f INPUT_FILE           Input fasta file for secretome prediction (REQUIRED)
  --run_name RUN_NAME, -n RUN_NAME            Prefix for output (default is input filename)
  --no_cleanup, -j                            Don't remove intermediate outputs
  --check, -x                                 Just check dependencies of secretome_pipe.py then exit (default: False)
  --verbose, -v                               Print verbose output (default: False)
  --permissive, -p                            Get permissive secretome prediction (default: False)
  --transporter_threshold TRANS, -t TRANS     Minimum number of tm domains in mature
                                              sequences required to consider a protein as a
                                              transporter

Requirements:

• linux (likely will work on MACOSX but is not tested)
• python3.4 or python2.7
• perl 5 (specifically installed at /usr/bin/perl)
• biopython 1.64

Run pip install -r requirements to ensure dependencies are installed.

You can then run tests to ensure the script and all the bundled dependencies work correctly by invoking ./test/test.py from the main directory.