def init_param_truth_file(truth_file, out_dir, hg_file):
# Calculate tumor purity and ploidy
# overall_ploidy = get_clone_ploidy(sorted_compl_cnv_file, hg_file, ["chrX", "chrY", "chrM"])
# Complement truth file with diploid regions
new_truth_file = os.path.join(out_dir,
"cntot_" + os.path.basename(truth_file))
with open(truth_file, "r") as tf:
tf_lines = tf.readlines()
if len(tf_lines[1].strip().split("\t")) == 5:
new_tf_lines = []
for tf_line in tf_lines:
(chr_id, start, end, cnA, cnB) = tf_line.strip().split("\t")
new_tf_lines.append("\t".join(
[chr_id, start, end,
str(int(cnA) + int(cnB))]))
with open(new_truth_file, "w") as new_tf:
new_tf.writelines("\n".join(new_tf_lines))
truth_file = new_truth_file
cnv_file = add_BED_complement(truth_file,
hg_file,
sort=True,
out_dir=out_dir,
hap_split=False)
#cnv_file = re.sub(".bed$", "_cnv.bed", out_file)
subprocess.Popen("rm %s" % re.sub("_sorted.bed", "_compl.bed", cnv_file),
shell=True)
if os.path.exists(new_truth_file):
subprocess.Popen("rm %s" % new_truth_file, shell=True)
return cnv_file
def init_param_truth_file(truth_file, out_dir, hg_file):
# Calculate tumor purity and ploidy
# overall_ploidy = get_clone_ploidy(sorted_compl_cnv_file, hg_file, ["chrX", "chrY", "chrM"])
# Complement truth file with diploid regions
new_truth_file = os.path.join(out_dir, "cntot_" + os.path.basename(truth_file))
with open(truth_file, "r") as tf:
tf_lines = tf.readlines()
if len(tf_lines[1].strip().split("\t")) == 5:
new_tf_lines = []
for tf_line in tf_lines:
(chr_id, start, end, cnA, cnB) = tf_line.strip().split("\t")
new_tf_lines.append("\t".join([chr_id, start, end, str(int(cnA)+int(cnB))]))
with open(new_truth_file, "w") as new_tf:
new_tf.writelines("\n".join(new_tf_lines))
truth_file = new_truth_file
cnv_file = add_BED_complement(truth_file, hg_file, sort=True, out_dir=out_dir, hap_split=False)
#cnv_file = re.sub(".bed$", "_cnv.bed", out_file)
subprocess.Popen("rm %s" % re.sub("_sorted.bed", "_compl.bed", cnv_file), shell=True)
if os.path.exists(new_truth_file):
subprocess.Popen("rm %s" % new_truth_file, shell=True)
return cnv_file
def main():
"""Evaluate Canvas results on simulated data"""
parser = OptionParser()
parser.add_option("-s", "--canvas_sim_id", dest="canvas_sim_id", type="string", help="simulation id (find Canvas results directories strating with sim_id")
parser.add_option("-d", "--canvas_sim_dir", dest="canvas_sim_dir", type="string", help="Canvas results directory")
(options, args) = parser.parse_args()
if options.canvas_sim_dir is None and options.canvas_sim_id is None:
parser.error("Specify Canvas results simulation id or directory")
if options.canvas_sim_dir is not None and options.canvas_sim_id is not None:
parser.error("Canvas results simulation id and directory are mutually exclusive options")
if options.canvas_sim_dir is not None:
res_dirs = [options.canvas_sim_dir]
if options.canvas_sim_id is not None:
canvas_dir = "/illumina/scratch/tmp/users/ccolombo/Canvas/"
res_dirs = [canvas_dir + cdir for cdir in os.listdir(canvas_dir) if os.path.isdir((canvas_dir + cdir)) and cdir.startswith(options.canvas_sim_id)]
if len(res_dirs) == 0:
print "No Canvas results directories found"
hg_file = "/home/ccolombo/filtered_human.hg19.genome"
for res_dir in res_dirs:
print res_dir
sim_dir = re.sub("/Canvas/", "/simulation/", res_dir)
res_id = re.sub("sim", "ev", res_dir.split("/")[-1])
if not os.path.exists(res_dir):
print("\nCanvas results directory %s does not exist\n\n" % res_dir)
continue
if not os.path.exists(res_dir):
print("\nSimulation directory %s does not exist\n\n" % sim_dir)
continue
if os.path.basename(res_dir).startswith("simNorm") or res_dir.endswith("_purity100"):
continue
out_dir = os.path.join("/illumina/scratch/tmp/users/ccolombo/evaluation/EvaluateCNV", res_id)
if not os.path.exists(out_dir):
os.mkdir(out_dir)
truth_file = create_het_cn_file(sim_dir, os.path.join(out_dir, "het_truth_file.bed"), no_dipl=True, hap_split=True, round=True)
#truth_file = "/illumina/scratch/tmp/users/ccolombo/simulation/HCC2218flt.bed"
truth_file = add_BED_complement(truth_file, hg_file, sort=False, out_dir=out_dir, hap_split=True)
perc_file = create_var_perc_file(sim_dir, os.path.join(out_dir, "var_perc.bed"), no_dipl=True)
# Write simulation parameters to file
out_file = os.path.join(out_dir, res_id + "_par.txt")
if not os.path.exists(out_file):
get_sim_params(sim_dir, truth_file, perc_file, out_file, hg_file)
# Run EvaluateCNV
out_file = os.path.join(out_dir, res_id + ".txt")
excl_file = "/illumina/scratch/tmp/users/ccolombo/evaluation/sim_filter.bed"
if not os.path.exists(out_file):
#if True:
evaluate_CNV(res_dir, truth_file, excl_file, out_file)
# Run EvaluateCNV only on heterogeneous variants
out_file = os.path.join(out_dir, res_id + "_onlyhet.txt")
#if True:
if not os.path.exists(out_file):
with open(excl_file, "r") as ef:
excl_vars = ef.readlines()
with open(perc_file, "r") as pf:
for line in pf:
(chr_id, start, end, cnA, cnB, perc) = line.strip().split("\t")
if float(perc) >= 0.8 and chr_id not in ["chrX", "chrY", "chrM"]:
excl_vars.append("\t".join([chr_id, start, end]) + "\n")
os.system("rm %s" % (perc_file))
excl_file = os.path.join(out_dir, "filter_onlyhet.bed")
with open(excl_file, "w") as wf:
wf.writelines(excl_vars)
evaluate_CNV(res_dir, truth_file, excl_file, out_file)
def main():
"""Evaluate Canvas results on simulated data"""
parser = OptionParser()
parser.add_option(
"-s",
"--canvas_sim_id",
dest="canvas_sim_id",
type="string",
help=
"simulation id (find Canvas results directories strating with sim_id")
parser.add_option("-d",
"--canvas_sim_dir",
dest="canvas_sim_dir",
type="string",
help="Canvas results directory")
(options, args) = parser.parse_args()
if options.canvas_sim_dir is None and options.canvas_sim_id is None:
parser.error("Specify Canvas results simulation id or directory")
if options.canvas_sim_dir is not None and options.canvas_sim_id is not None:
parser.error(
"Canvas results simulation id and directory are mutually exclusive options"
)
if options.canvas_sim_dir is not None:
res_dirs = [options.canvas_sim_dir]
if options.canvas_sim_id is not None:
canvas_dir = "/illumina/scratch/tmp/users/ccolombo/Canvas/"
res_dirs = [
canvas_dir + cdir for cdir in os.listdir(canvas_dir)
if os.path.isdir((canvas_dir +
cdir)) and cdir.startswith(options.canvas_sim_id)
]
if len(res_dirs) == 0:
print "No Canvas results directories found"
hg_file = "/home/ccolombo/filtered_human.hg19.genome"
for res_dir in res_dirs:
print res_dir
sim_dir = re.sub("/Canvas/", "/simulation/", res_dir)
res_id = re.sub("sim", "ev", res_dir.split("/")[-1])
if not os.path.exists(res_dir):
print("\nCanvas results directory %s does not exist\n\n" % res_dir)
continue
if not os.path.exists(res_dir):
print("\nSimulation directory %s does not exist\n\n" % sim_dir)
continue
if os.path.basename(res_dir).startswith("simNorm") or res_dir.endswith(
"_purity100"):
continue
out_dir = os.path.join(
"/illumina/scratch/tmp/users/ccolombo/evaluation/EvaluateCNV",
res_id)
if not os.path.exists(out_dir):
os.mkdir(out_dir)
truth_file = create_het_cn_file(sim_dir,
os.path.join(out_dir,
"het_truth_file.bed"),
no_dipl=True,
hap_split=True,
round=True)
#truth_file = "/illumina/scratch/tmp/users/ccolombo/simulation/HCC2218flt.bed"
truth_file = add_BED_complement(truth_file,
hg_file,
sort=False,
out_dir=out_dir,
hap_split=True)
perc_file = create_var_perc_file(sim_dir,
os.path.join(out_dir, "var_perc.bed"),
no_dipl=True)
# Write simulation parameters to file
out_file = os.path.join(out_dir, res_id + "_par.txt")
if not os.path.exists(out_file):
get_sim_params(sim_dir, truth_file, perc_file, out_file, hg_file)
# Run EvaluateCNV
out_file = os.path.join(out_dir, res_id + ".txt")
excl_file = "/illumina/scratch/tmp/users/ccolombo/evaluation/sim_filter.bed"
if not os.path.exists(out_file):
#if True:
evaluate_CNV(res_dir, truth_file, excl_file, out_file)
# Run EvaluateCNV only on heterogeneous variants
out_file = os.path.join(out_dir, res_id + "_onlyhet.txt")
#if True:
if not os.path.exists(out_file):
with open(excl_file, "r") as ef:
excl_vars = ef.readlines()
with open(perc_file, "r") as pf:
for line in pf:
(chr_id, start, end, cnA, cnB,
perc) = line.strip().split("\t")
if float(perc) >= 0.8 and chr_id not in [
"chrX", "chrY", "chrM"
]:
excl_vars.append("\t".join([chr_id, start, end]) +
"\n")
os.system("rm %s" % (perc_file))
excl_file = os.path.join(out_dir, "filter_onlyhet.bed")
with open(excl_file, "w") as wf:
wf.writelines(excl_vars)
evaluate_CNV(res_dir, truth_file, excl_file, out_file)