def build_qsub(self):
"""
Builds and writes this CallObject's qsub script to current working directory
using options provided under the "qsub_options" sub-tree in the yaml config file.
"""
nicknames = {'tophat':'th',
'cufflinks':'cl',
'cuffmerge':'cm',
'cuffdiff':'cd',}
qsub_options = self.yargs.qsub_options
# set keyword args for template
kw = Bunch()
kw.queues = qsub_options.queues
kw.datahome = qsub_options.datahome
kw.core_range = qsub_options.core_range
kw.email_addy = self.email_info.email_to
kw.call_id = self.call_id
job_name = "%s_%s" % (nicknames[self.prog_name], '_'.join(self.call_id.split('_')[1:]))
kw.job_name = job_name
kw.out_dir = self.out_dir
kw.ld_library_path = qsub_options.ld_library_path
# need to make sure we use the number of cores that the SGE gave us
kw.cmd_str = self.cmd_string.replace('-p %s' % (self.opt_dict['p']),'-p $CORES')
qsub_template = Template(filename=qsub_options.template)
out_file = open('%s.qsub.sh' % (self.call_id),'w')
qsub_string = qsub_template.render(**kw)
out_file.write(qsub_string)
out_file.close()
def main():
"""
The main loop. Lets ROCK!
"""
desc = """This script reads the files in a cuffdiff output directory into cummeRbund, generates some standard preliminary plots, and saves the output."""
parser = argparse.ArgumentParser(description=desc)
parser.add_argument('--version', action='version', version='%(prog)s ' + blacktie.__version__,
help="""Print version number.""")
parser.add_argument('--cuffdiff-dir', type=str,
help="""Path to a cuffdiff output directory.""")
#parser.add_argument('--cummerbund-db', type=str,
#help="""Path to a pre-built cummeRbund 'cuffData.db'. (this is rarely specified directly; usually --cuffdiff-dir works fine)""")
parser.add_argument('--gtf-path', type=str, default='NULL',
help="""Path to gtf file used in cuffdiff analysis. This will provide transcript model information.""")
parser.add_argument('--genome', type=str, default='NULL',
help="""String indicating which genome build the .gtf annotations are for (e.g. 'hg19' or 'mm9').""")
parser.add_argument('--out', type=str,
help="""A base directory to add to our saved plots into.""")
parser.add_argument('--file-type', type=str, choices=['pdf','jpeg','png','ps'], default='pdf',
help="""The type of output file to use when saving our plots. (default: %(default)s)""")
if len(sys.argv) == 1:
parser.print_help()
exit(0)
args = parser.parse_args()
# import the cummeRbund libray to the R workspace
import_cummeRbund_library()
# read in the cuffdiff data
cuff = r.readCufflinks(dir=args.cuffdiff_dir, gtfFile=args.gtf_path, genome=args.genome)
# Find out if we have replicates
genes_rep_fpkm = r.repFpkm(r.genes(cuff))
replicate_ids = set(genes_rep_fpkm[2])
if len(replicate_ids) > 1:
we_have_replicates = True
else:
we_have_replicates = False
# Store my plots here
rplots = Bunch()
# dispersion plot
rplots.dispersionPlot = r.dispersionPlot(r.genes(cuff))
# SCV plots
if we_have_replicates:
rplots.fpkmSCVPlot_genes = r.fpkmSCVPlot(r.genes(cuff))
rplots.fpkmSCVPlot_isoforms = r.fpkmSCVPlot(r.isoforms(cuff))
else:
pass
# Density Plots
rplots.csDensity = r.csDensity(r.genes(cuff))
if we_have_replicates:
rplots.csDensity_reps = r.csDensity(r.genes(cuff),replicates='T')
else:
pass
# Box Plots
rplots.csBoxplot = r.csBoxplot(r.genes(cuff))
if we_have_replicates:
rplots.csBoxplot_reps = r.csBoxplot(r.genes(cuff),replicates='T')
else:
pass
# Scatter Matrix
rplots.csScatterMatrix = r.csScatterMatrix(r.genes(cuff))
# TODO: csDendro does not use ggplot2 it seems so ggsave() does not work. When issue is fixed, uncomment this.
## # Dendrograms
## rplots.csDendro = r.csDendro(r.genes(cuff))
##
## if we_have_replicates:
## rplots.csDendro_reps = r.csDendro(r.genes(cuff),replicates='T')
# Volcano Matrix
rplots.csVolcanoMatrix = r.csVolcanoMatrix(r.genes(cuff))
# Sig Matrix
rplots.sigMatrix = r.sigMatrix(cuff,level='genes',alpha=0.05)
# get significant genes
mySigGeneIds = r.getSig(cuff,alpha=0.05,level='genes')
mySigGenes = r.getGenes(cuff,mySigGeneIds)
print "Significant Genes: %s" % (len(mySigGeneIds))
# Preliminary Clustering
ic = r.csCluster(mySigGenes,k=20)
rplots.csClusterPlot = r.csClusterPlot(ic)
# print the plots
print_my_plots(r, rplots, out=args.out, file_type=args.file_type)