run_prefab.py.README

Wrapper for running and scoring alignment programs on Prefab

The two important files are run_prefab.py (the script itself) and
prefab4_file_list.txt, which contains a list of all prefab input
files, their number of sequences and (average, pairwise) sequence
identity.

Download prefab from http://www.drive5.com/muscle/prefab.htm (and
qscore from http://www.drive5.com/qscore/). See PREFAB section in
http://www.ncbi.nlm.nih.gov/pubmed/15034147: Edgar RC. MUSCLE:
multiple sequence alignment with high accuracy and high throughput.
Nucleic Acids Res. 2004 Mar 19;32(5):1792-7.



Creating prefab4_file_list.txt
==============================

Assuming you have squid's alistat installed:

<code>
d=./prefab4
for fbase in $(find $d/ref -type f -name [0-9]*_[0-9]* | xargs -n1 basename); do
   pwidref=$(alistat $d/ref/$fbase | awk '/^Average identity:/ {sub(/%/, ""); print $NF}');
   numseqin=$(grep -c -- '^>' $d/in/$fbase);
   echo "$fbase $numseqin $pwidref";
done > ${d}_file_list.txt
</code>



Running an aligner (run_prefab.py)
==================================

Call the wrapper like this:

$ ./run_prefab.py \
 -c "myaligner -in @IN@ -out @OUT@ -more-options -even-more-options" \
 -o outputdir -v

This would run the program "myaligner" on all of Prefab (replacing
@IN@ and @OUT@ accordingly) and store resulting alignments and qscore
files in "outputdir"

To run vanilla Muscle you'd use:
$ ./run_prefab.py \
 -c "your-path-to-muscle -in @IN@ -out @OUT@" \
 -o muscle_out -v

To run ClustalO with hmm-iteration:
$ ./run_prefab.py \
 -c "your-path-to-clustalo -i @IN@ -o @OUT@ --hi 1" \
 -o clustalo_hi-1 -v


Existing files will never be overwritten, instead an error message
will be printed to stderr. Qscore files will be called
outputdir/prefabid/prefabid.out_qscore.txt. It's up to you to analyse
the score files afterwards (use prefab4_file_list.txt if you want to
limit yourself to certain id-ranges). A quick and dirty approach to
calculate the average qscore would be:

$ find outdir -name  \*out_qscore.txt -exec cat {} \; | \
  awk -F';' '{sub(/Q=/, ""); s+=$3; n+=1}
  END {printf "avg. qscore (%d files) = %f\n", n, s/n}'


The script allows you to limit the runs to files with a certain
number of sequences (--min-nseq/--max-nseq) or a certain
identity-range (--min-id/--max-id).

In doubt try ./run_prefab.py  -h