def run(self):
ensureDir(self.cfg.gene_set_dir)
ensureDir(self.cfg.metrics_dir)
generate_gene_set_wrapper(self.cfg)
for gp, gtf in zip(*[self.cfg.geneset_gps.itervalues(), self.cfg.geneset_gtfs.itervalues()]):
self.convert_gp_to_gtf(gp, gtf)
self.convert_gp_to_gtf(self.cfg.combined_gp, self.cfg.combined_gtf)
def main():
args = parse_args()
if args.name is None:
args.name = args.database
ensureDir(args.outDir)
get_genes(args.database, args.name, args.outDir, args.includeChroms, args.convertUCSCtoEnsembl)
build_attributes(args.database, args.name, args.outDir)
def init_image(path, width, height):
"""
Sets up a PDF object.
"""
ensureDir(os.path.dirname(path))
pdf = plt_back.PdfPages(path)
# width by height in inches
fig = plt.figure(figsize=(width, height), dpi=300, facecolor='w')
return fig, pdf
def run(self):
for biotype, tm_cfg in self.cfg.tm_plots.iteritems():
ensureDir(tm_cfg.output_dir)
paralogy_plot(tm_cfg.ordered_target_genomes, tm_cfg.query_genome, biotype, tm_cfg.para_plot,
self.cfg.db)
cov_plot(tm_cfg.ordered_target_genomes, tm_cfg.query_genome, biotype, tm_cfg.cov_plot, self.cfg.db)
ident_plot(tm_cfg.ordered_target_genomes, tm_cfg.query_genome, biotype, tm_cfg.ident_plot, self.cfg.db)
num_pass_excel(tm_cfg.ordered_target_genomes, tm_cfg.query_genome, biotype, tm_cfg.num_pass_excel,
self.cfg.db, self.cfg.args.filterChroms)
num_pass_excel_gene_level(tm_cfg.ordered_target_genomes, tm_cfg.query_genome, biotype,
tm_cfg.num_pass_excel_gene, self.cfg.db, self.cfg.args.filterChroms)
def start_jobtree(self, args, entry_fn, norestart=False):
"""
Start a jobTree. Based on the flag norestart, will decide if we should attempt a restart.
TODO: this hack re-creates the namespace to avoid import issues.
"""
tmp_cfg = argparse.Namespace()
tmp_cfg.__dict__.update(vars(args))
ensureDir(os.path.dirname(args.jobTree))
jobtree_path = args.jobTree
if norestart is True or not os.path.exists(jobtree_path) or self.jobtree_is_finished(jobtree_path):
self.make_jobtree_dir(jobtree_path)
entry_fn(tmp_cfg)
else: # try restarting the tree
try:
entry_fn(args)
except RuntimeError: # try starting over
self.make_jobtree_dir(jobtree_path)
entry_fn(tmp_cfg)
def main():
args = parse_args()
data = defaultdict(lambda: defaultdict(lambda: defaultdict(list)))
ref_path = os.path.join(args.ref_dir, args.experiment)
tgt_path = os.path.join(args.tgt_dir, args.experiment)
if args.genomes is None:
args.genomes = os.listdir(ref_path)
common_name_map = get_common_name_map(args.attrs)
load_dir(data, 'ref', ref_path, args.genomes, common_name_map)
load_dir(data, 'tgt', tgt_path, args.genomes, common_name_map)
base_counts_dir = os.path.join(args.counts_dir, args.experiment)
construct_counts_tables(data, base_counts_dir)
deseq_path = os.path.join(args.deseq_dir, args.experiment)
ensureDir(deseq_path)
for g in args.genomes:
runProc(['Rscript', 'run_DEseq.R', g, deseq_path, base_counts_dir])
out_dir = os.path.join(args.out_dir, args.experiment)
generate_plots(args.genomes, deseq_path, out_dir)
def __init__(self, paraHost, runDir, paraDir, jobFile=None, cpu=None, mem=None, maxJobs=None, retries=None):
""""will chdir to run dir.. paraDir should be relative
to runDir or absolute, jobFile should be relative to runDir
or absolute.
"""
self.paraHost = paraHost
# symlinks can confuse parasol, as it can give two different names for a job.
self.runDir = os.path.realpath(os.path.abspath(runDir))
self.paraDir = os.path.realpath(paraDir)
self.jobFile = jobFile
self.cpu = cpu
self.mem = mem
self.maxJobs = maxJobs
self.retries = retries
fileOps.ensureDir(self._mkAbs(self.runDir, self.paraDir))
if jobFile is not None:
absJobFile = self._mkAbs(self.runDir, self.jobFile)
if not os.path.exists(absJobFile):
raise PycbioException("job file not found: {}".format(absJobFile))
def generate_plots(genomes, deseq_path, out_dir):
for g in genomes:
base_out = os.path.join(out_dir, g)
ensureDir(base_out)
main_plot = os.path.join(base_out, 'nonzero_log.pdf')
para_plot = os.path.join(base_out, 'nonzero_log_paralogy.pdf')
ref_col_name = 'Normalized counts (aligned to mm10)'
tgt_col_name = 'Normalized counts (aligned to {})'.format(g)
df = pd.read_table(os.path.join(deseq_path, g + '.data_table.tsv'), sep=' ', header=0, index_col=0)
df = df[df['ref'] >= 1]
df = df[df['tgt'] >= 1]
df = strip_version(df)
df = calculate_pval_vector(df, ref_col_name, tgt_col_name)
p = sns.lmplot(x=ref_col_name, y=tgt_col_name, data=df, scatter_kws={"s": 6, "alpha": 0.7}, hue='Adjusted p-value', fit_reg=False)
p.set(xlim=(0, 10), ylim=(0, 10))
p.savefig(main_plot, format='pdf')
df2 = load_paralogy(g, df)
p2 = sns.lmplot(x=ref_col_name, y=tgt_col_name, data=df2, scatter_kws={"s": 6, "alpha": 0.7}, col='Paralogy', hue='Adjusted p-value', fit_reg=False)
p2.set(xlim=(0, 10), ylim=(0, 10))
p2.savefig(para_plot, format='pdf')
plt.close('all')
def write(self, outDir):
fileOps.ensureDir(outDir)
frame = self.__mkFrame(self.title, self.dirPercent, self.below)
frame.writeFile(outDir + "/index.html")
self.__writeDirPages(outDir)
def getOutputDir(self):
"""get the path to the output directory to use for this test, create if it doesn't exist"""
d = self.getTestDir() + "/output";
fileOps.ensureDir(d)
return d
def getOutputDir(self):
"""get the path to the output directory to use for this test, create if it doesn't exist"""
d = self.getTestDir() + "/output"
fileOps.ensureDir(d)
return d
def run(self):
ensureDir(self.cfg.chaining.out_dir)
self.start_jobtree(self.cfg.chaining, chainNetStartup, norestart=self.cfg.args.norestart)
def build_out_dirs(out_dir, genome, institute, tissue, experiment):
out_path = os.path.join(out_dir, genome, institute, tissue, experiment)
ensureDir(out_path)
return out_path
