I'm merely the installer of SGA on a shared HPC resource, but I wanted to
say thanks for providing some examples so that an installation can be tested
out, ahead of passing it over for use by the bioinformaticians, however , I
did get a little confused.
The sga-celegans.sh script hard-wires a version (now old) of the main SGA
binary in a variable, SGA_BIN, which is the used throughout the script, vis:
# Parameters
SGA_BIN=sga-0.9.31
however a make install doesn't see a versioned-binary installed, just a bare sga.
Later on in the script, there's a couple of hardcoded references to paths
to other SGA scripts, with non-versioned names, that SGS does install (though see below).
# Realign reads to the contigs
~/work/devel/sga/src/bin/sga-align --name celegans.pe $CTGS $IN1 $IN2
# Make contig-contig distance estimates
~/work/devel/sga/src/bin/sga-bam2de.pl -n $MIN_PAIRS --prefix libPE celegans.pe.bam
# Make contig copy number estimates
~/work/devel/sga/src/bin/sga-astat.py -m $MIN_LENGTH celegans.pe.refsort.bam > libPE.astat
I note that in the sga-ecoli-miseq.sh script, everything, including the scripts above, has
been turned into a variable
so I wondered if you might update the sga-celegans.sh script in a similar manner.
Having resolved that confusion, I then noticed that the make install I had done, had
only placed these four files
sga sga-astat.py sga-bam2de.pl sga-mergeDriver.pl
into the bin directory and not copied over the sga-align script that the sga-celegans.sh
example script refers to, indeed there are a number of scripts in the source distribution's
src/bin directory that are not installed by a make install from the src dir.
Apologies if I have missed something obvious.
I am working with a 0.10.14 tarball.
I'm merely the installer of SGA on a shared HPC resource, but I wanted to say thanks for providing some examples so that an installation can be tested out, ahead of passing it over for use by the bioinformaticians, however , I did get a little confused.
The
sga-celegans.shscript hard-wires a version (now old) of the main SGA binary in a variable,SGA_BIN, which is the used throughout the script, vis:however a
make installdoesn't see a versioned-binary installed, just a baresga.Later on in the script, there's a couple of hardcoded references to paths to other SGA scripts, with non-versioned names, that SGS does install (though see below).
I note that in the
sga-ecoli-miseq.shscript, everything, including the scripts above, has been turned into a variableso I wondered if you might update the
sga-celegans.shscript in a similar manner.Having resolved that confusion, I then noticed that the
make installI had done, had only placed these four filesinto the
bindirectory and not copied over thesga-alignscript that thesga-celegans.shexample script refers to, indeed there are a number of scripts in the source distribution'ssrc/bindirectory that are not installed by amake installfrom thesrcdir.Apologies if I have missed something obvious. I am working with a 0.10.14 tarball.