def main():
parser = argparse.ArgumentParser(
description="Calculate TransFIC labels")
cmd = Command.withtraits(DbTrait, PredictorsInDbTrait)(parser)
cmd.add_db_args()
parser.add_argument("cutoffs_path", metavar="CUTOFFS",
help="File containing the cutoffs")
cmd.add_selected_predictors_args()
parser.add_argument("-u", "--updated-predictors", dest="updated_predictors", metavar="NAMES",
help="Updated predictor names")
args, logger = cmd.parse_args("calc-label")
db = cmd.open_db()
try:
logger.info("Loading state ...")
state = load_weights(args.cutoffs_path)
avail_predictors, stats = [state[k] for k in ["predictors", "stats"]]
predictors = cmd.get_selected_predictors(default_all=True)
missing_predictors = [p for p in predictors if p not in set(avail_predictors)]
if len(missing_predictors) > 0:
raise Exception("Missing cutoff stats for predictors: {}".format(", ".join(missing_predictors)))
if args.updated_predictors is not None:
if len(predictors) != len(args.updated_predictors):
raise Exception("The number of selected predictors does not match the number of predictor names to update")
updated_predictors = dict([(p, u) for p, u in zip(predictors, args.updated_predictors)])
else:
updated_predictors = dict([(p, "{}_LABEL".format(p)) for p in predictors])
# create predictors in the database if required
db_predictors = set([p["id"] for p in db.predictors()])
for predictor, updated_predictor in updated_predictors.items():
if updated_predictor not in db_predictors:
logger.info("Creating predictor {} ...".format(updated_predictor))
db.add_predictor(updated_predictor, FannsDb.CALCULATED_PREDICTOR_TYPE, source=[predictor])
cutoffs = {}
for predictor in predictors:
cutoff_low_mid, cutoff_mid_high = [stats[predictor][v] for v in ["cutoff_low_mid", "cutoff_mid_high"]]
logger.info("{}: cutoffs: low_mid={}, mid_high={}".format(predictor, cutoff_low_mid, cutoff_mid_high))
cutoffs[predictor] = (cutoff_low_mid, cutoff_mid_high)
logger.info("Calculating ...")
progress = RatedProgress(logger, name="SNVs")
for num_rows, row in enumerate(db.query_scores(predictors=predictors), start=1):
scores = row["scores"]
uscores = {}
for predictor in predictors:
score = scores[predictor]
if score is None:
continue
cutoff_low_mid, cutoff_mid_high = cutoffs[predictor]
updated_predictor = updated_predictors[predictor]
uscores[updated_predictor] = 0.0 if score < cutoff_low_mid else 1.0 if score < cutoff_mid_high else 2.0
if len(uscores) > 0:
db.update_scores(row["id"], uscores)
progress.update()
db.commit()
except:
cmd.handle_error()
return 0
def main():
parser = argparse.ArgumentParser(
description="Import scores into the database")
cmd = Command.withtraits(DbTrait, PredictorsTrait)(parser)
cmd.add_db_args()
parser.add_argument("source_path", metavar="SOURCE",
help="The source file. Use - for standard input.")
#TODO: which are the coordinates column
cmd.add_selected_predictors_args()
parser.add_argument("--skip-empty-scores", dest="skip_empty_scores", action="store_true", default=False,
help="Skip SNV's where all the scores are empty")
parser.add_argument("--skip-update-predictors", dest="skip_update_predictors", action="store_true", default=False,
help="Skip the update of the predictors.")
parser.add_argument("--skip-create-index", dest="skip_create_index", action="store_true", default=False,
help="Skip the creation of the database indices.")
parser.add_argument("--ignore-errors", dest="ignore_errors", action="store_true", default=False,
help="When errors on the input file, report them but continue processing the input.")
args, logger = cmd.parse_args("import")
db = cmd.open_db()
try:
progress = RatedProgress(logger, name="SNVs")
total_lines = 0
logger.info("Reading {} ...".format(args.source_path if args.source_path != "-" else "from standard input"))
with tsv.open(args.source_path) as f:
# Parse header
hdr_line = f.readline()
hdr = {}
for index, name in enumerate(hdr_line.rstrip("\n").split("\t")):
hdr[name] = index
# Predictors to update from the user selection and source availability
db_predictors = set([p["id"] for p in db.predictors()])
src_predictors = [name for name in hdr if name not in COORD_COLUMNS]
predictors = cmd.get_selected_predictors(available_predictors=src_predictors)
for predictor in predictors:
if predictor not in db_predictors:
logger.info("Creating non existing predictor: {}".format(predictor))
db.add_predictor(predictor, FannsDb.SOURCE_PREDICTOR_TYPE)
logger.info("Predictors: {}".format(", ".join(predictors)))
all_columns = COORD_COLUMNS + predictors
types = COORD_TYPES + ([score_value] * len(predictors))
missing_columns = [name for name in all_columns if name not in hdr]
if len(missing_columns) > 0:
raise Exception("The following columns are missing: {}".format(", ".join(missing_columns)))
columns = [hdr[name] for name in all_columns]
max_column = max(columns)
for line_num, line in enumerate(f, start=2):
fields = line.rstrip("\n").split("\t")
if len(fields) < max_column:
log.error("Missing columns for line {}: {}".format(line_num, " ".join(fields)))
if not args.ignore_errors:
raise
try:
fields = [type_cast(fields[index]) for type_cast, index in zip(types, columns)]
except Exception as ex:
logger.error("{} at line {}: {}".format(str(ex), line_num, " ".join(fields)))
if not args.ignore_errors:
raise
(chr, strand, start, ref, alt, transcript,
aa_pos, aa_ref, aa_alt, protein) = fields[:10]
scores = fields[10:]
if args.skip_empty_scores and sum([0 if s is None else 1 for s in scores]) == 0:
continue
try:
db.add_snv(
chr=chr, strand=strand, start=start, ref=ref, alt=alt, transcript=transcript,
protein=protein, aa_pos=aa_pos, aa_ref=aa_ref, aa_alt=aa_alt,
scores=dict(zip(predictors, scores)))
except Exception as ex:
logger.error("Error importing SNV at line {}: {}".format(line_num, str(ex)))
if not args.ignore_errors:
raise
progress.update()
total_lines += line_num
progress.log_totals()
logger.info("Finalizing database ...")
if not args.skip_update_predictors:
logger.info("Updating predictors ...")
db.update_predictors()
logger.info("Committing ...")
db.commit()
if not args.skip_create_index:
logger.info("Creating indices ...")
db.create_indices()
logger.info("Finished successfully. Elapsed time: {}".format(progress.elapsed_time))
except:
return cmd.handle_error()
finally:
db.close()
return 0
def main():
parser = argparse.ArgumentParser(description="Calculate TransFIC for the selected scores")
cmd = Command.withtraits(DbTrait, PredictorsInDbTrait, TransformsTrait)(parser)
cmd.add_db_args()
parser.add_argument(
"feature_name",
metavar="FEATURE_COLUMN",
help="The column name with the features. It can be transcript, protein or any of the available annotations.",
)
parser.add_argument("blt_path", metavar="BLT_PATH", help="The baseline tolerance statistics.")
cmd.add_selected_predictors_args()
parser.add_argument(
"-u", "--updated-predictors", dest="updated_predictors", metavar="NAME", help="Updated predictor names"
)
cmd.add_transform_args()
args, logger = cmd.parse_args("calc")
db = cmd.open_db()
# initialize feature selection
db_annotations = [a["id"] for a in db.maps()]
if args.feature_name not in set(["transcript", "protein"] + db_annotations):
logger.error("Feature name not available in the database: {}".format(args.feature_name))
logger.error("Available annotations: {}".format(", ".join(db_annotations)))
exit(-1)
if args.feature_name.lower() in ["transcript", "protein"]:
annotations = None
feature_getter = lambda row: row[args.feature_name]
else:
annotations = [args.feature_name]
feature_getter = lambda row: row["annotations"][args.feature_name]
# predictors, transforms, and updated_predictors
predictors = cmd.get_selected_predictors(default_all=True)
transforms = cmd.get_transforms()
if args.updated_predictors is not None:
if len(predictors) != len(args.updated_predictors):
logger.error("The number of selected predictors does not match the number of predictor names to update")
exit(-1)
updated_predictors = dict([(p, u) for p, u in zip(predictors, args.updated_predictors)])
else:
updated_predictors = dict([(p, "TFIC_{}".format(p)) for p in predictors])
# create predictors in the database if required
db_predictors = set([p["id"] for p in db.predictors()])
for predictor, updated_predictor in updated_predictors.items():
if updated_predictor not in db_predictors:
logger.info("Creating predictor {} ...".format(updated_predictor))
db.add_predictor(updated_predictor, FannsDb.CALCULATED_PREDICTOR_TYPE, source=[predictor])
try:
logger.info("Loading baseline tolerance statistics ...")
with tsv.open(args.blt_path) as f:
doc = json.load(f)
blt_predictors = doc["predictors"]
features = doc["features"]
blt_stats = doc["blt"]
num_predictors = len(blt_predictors)
logger.info(" Predictors: {}".format(", ".join(blt_predictors)))
logger.info(" Features: {}".format(len(features)))
logger.info("Calculating ...")
progress = RatedProgress(logger, name="SNVs")
rows_count = updated_count = 0
for row in db.query_scores(predictors=predictors, maps=annotations):
rows_count += 1
scores = row["scores"]
feature = feature_getter(row)
if feature not in blt_stats:
continue
feature_stats = blt_stats[feature]
tfic_scores = calculate_tfic(predictors, updated_predictors, feature_stats, scores, transforms)
if len(tfic_scores) > 0:
db.update_scores(row["id"], tfic_scores)
updated_count += 1
progress.update()
progress.log_totals()
logger.info("Commit ...")
db.commit()
logger.info(
"Finished. Total rows = {}, updated rows = {}, elapsed_time = {}".format(
rows_count, updated_count, progress.elapsed_time
)
)
except:
cmd.handle_error()
return 0
import os from bgcore import logging as bglogging from fannsdb.cmdhelper import Command, PredictorsTrait, TransformsTrait DefaultCommandHelper = Command.withtraits(PredictorsTrait, TransformsTrait)
def main():
parser = argparse.ArgumentParser(
description="Calculate TransFIC for the selected scores")
cmd = Command(parser)
parser.add_argument("weights", metavar="WEIGHTS_PATH", nargs="+",
help="The list of files containing weights.")
parser.add_argument("-c", dest="comparisons", metavar="S1/S2", action="append",
help="Compare score S1 with S2")
parser.add_argument("-o", dest="out_path", metavar="PATH",
help="The path to save the plot image.")
parser.add_argument("-W", "--width", dest="fig_width", metavar="WIDTH", type=int,
help="The image width.")
parser.add_argument("-H", "--height", dest="fig_height", metavar="HEIGHT", type=int,
help="The image height.")
parser.add_argument("--dpi", dest="fig_dpi", metavar="DPI", type=int, default=100,
help="The image dpi.")
parser.add_argument("-i", "--interactive", dest="interactive", action="store_true", default=False,
help="Show the plot in interactive mode.")
parser.add_argument("-t", "--title", dest="title", metavar="TITLE",
help="The plot title.")
args, logger = cmd.parse_args("perf-plot")
try:
width = 0.5
if args.comparisons is None:
raise Exception("Select the predictors to compare with -c please.")
comparisons = []
for c in args.comparisons:
s = c.split("/")
if len(s) != 2:
raise Exception("Wrong comparison format: {}".format(c))
comparisons += [tuple(s)]
num_comparisons = len(comparisons)
if not args.interactive:
import matplotlib
matplotlib.use('Agg')
from matplotlib import pylab
#pylab.rcParams['xtick.major.pad'] = '8'
pylab.rcParams['ytick.major.pad'] = '20'
fig = plt.figure(figsize=(args.fig_width or 12, args.fig_height or 10.4), dpi=args.fig_dpi or 100)
fig.subplots_adjust(left=0.11, bottom=0.03, right=0.99, top=0.94, wspace=0.07, hspace=0.53)
if args.title is not None:
fig.suptitle(args.title)
num_rows = len(args.weights)
for row_idx, wpath in enumerate(args.weights, start=0):
basename = os.path.basename(wpath)
name = basename
if name.endswith(".json"):
name = name[:-5]
if name.endswith("-weights"):
name = name[:-8]
i = name.find("__")
if i > 0:
name1, name2 = name.split("__")[:2]
else:
name1 = name2 = name
logger.info("{} ...".format(name))
with open(wpath) as f:
state = json.load(f)
metrics = state["metrics"]
x = np.arange(0, 100, 10)
y = np.arange(2)
j = 0
for p1, p2 in comparisons:
if p1 not in metrics:
raise Exception("Predictor '{}' not found in statistics file {}".format(p1, basename))
if p2 not in metrics:
raise Exception("Predictor '{}' not found in statistics file {}".format(p2, basename))
stats1 = metrics[p1]
stats2 = metrics[p2]
mcc1 = stats1["best_perf"]["MCC"] * 100.0
mcc2 = stats2["best_perf"]["MCC"] * 100.0
row_title = "{}\n{}".format(name1, name2)
col_title = p1 if row_idx == 0 else ""
ax = fig.add_subplot(num_rows, num_comparisons, row_idx * num_comparisons + j + 1, title=col_title)
ax.set_xticks(x)
ax.set_xlim(0, 100)
ax.xaxis.grid(True)
ax.set_yticks([0.25, 0.75])
if j == 0:
ax.set_ylabel(row_title, rotation=90)
ax.set_yticklabels(["Orig", "TFIC"])
else:
ax.set_yticklabels(["", ""])
b1 = ax.barh([0.0], [mcc1], width, color="r")
b2 = ax.barh([0.5], [mcc2], width, color="y")
j += 1
#ax.set_xticklabels(tuple([p1 for p1, p2 in comparisons]))
#ax.legend((b1[0], b2[0]), ("Original score", "TransFIC score"))
if args.out_path is not None:
plt.savefig(args.out_path, bbox_inches=0)
if args.interactive:
plt.show()
except:
cmd.handle_error()
return 0
