def parse_mod_data(
res_dir,
out_fp,
valid_sites,
include_strand,
samp_lab,
max_stats,
ctrl_sites=None,
):
mod_acc, parsim_acc, aligned_lens = report_acc_metrics(
res_dir, out_fp, samp_lab)
ctrl_data = None
mods_db_fn = mh.get_megalodon_fn(res_dir, mh.PR_MOD_NAME)
if os.path.exists(mods_db_fn):
if ctrl_sites is not None:
all_site_stats = mods.extract_stats_at_valid_sites(
mods_db_fn,
valid_sites + ctrl_sites,
include_strand=include_strand,
max_stats=max_stats,
)
mods_data = all_site_stats[:len(valid_sites)]
ctrl_data = all_site_stats[len(valid_sites):]
elif valid_sites is not None:
mods_data = mods.extract_stats_at_valid_sites(
mods_db_fn,
valid_sites,
include_strand=include_strand,
max_stats=max_stats,
)
else:
mods_data = [
mods.extract_all_stats(mods_db_fn, max_stats=max_stats),
]
else:
mods_data = None
return VAL_MOD_DATA(mod_acc, parsim_acc, aligned_lens, mods_data,
ctrl_data, samp_lab)
def parse_mod_data(
res_dir, out_fp, valid_sites, include_strand, samp_lab,
ctrl_sites=None):
mod_acc, parsim_acc = report_acc_metrics(res_dir, out_fp, samp_lab)
ctrl_data = None
mods_db_fn = mh.get_megalodon_fn(res_dir, mh.PR_MOD_NAME)
try:
if ctrl_sites is not None:
all_site_stats = mods.extract_stats_at_valid_sites(
mods_db_fn, valid_sites + ctrl_sites,
include_strand=include_strand)
mods_data = all_site_stats[:len(valid_sites)]
ctrl_data = all_site_stats[len(valid_sites):]
elif valid_sites is not None:
mods_data = mods.extract_stats_at_valid_sites(
mods_db_fn, valid_sites, include_strand=include_strand)
else:
mods_data = [mods.extract_all_stats(mods_db_fn), ]
except (FileNotFoundError, mh.MegaError):
mods_data = None
return VAL_MOD_DATA(mod_acc, parsim_acc, mods_data, ctrl_data, samp_lab)
def _main(args):
logging.init_logger(quiet=args.quiet)
if (args.ground_truth_data is None
and args.control_megalodon_results_dir is None):
LOGGER.error(
"Must provide either --control-megalodon-results-dir or " +
"--ground-truth-data")
sys.exit()
db_fn = mh.get_megalodon_fn(args.megalodon_results_dir, mh.PR_MOD_NAME)
if args.ground_truth_data is not None:
LOGGER.info("Parsing ground truth data")
gt_mod_pos, gt_can_pos = mh.parse_ground_truth_file(
args.ground_truth_data, include_strand=args.strand_specific_sites)
LOGGER.info(
("Loaded ground truth data with {} modified sites and {} " +
"canonical sites.").format(len(gt_mod_pos), len(gt_can_pos)))
LOGGER.info("Reading ground truth modified base statistics from " +
"database.")
all_mod_llrs, all_can_llrs = mods.extract_stats_at_valid_sites(
db_fn,
[gt_mod_pos, gt_can_pos],
include_strand=args.strand_specific_sites,
)
else:
LOGGER.info("Reading ground truth modified base statistics from " +
"database")
all_mod_llrs = mods.extract_all_stats(db_fn)
LOGGER.info("Reading ground truth modified base statistics from " +
"canonical sample database")
all_can_llrs = mods.extract_all_stats(
mh.get_megalodon_fn(args.control_megalodon_results_dir,
mh.PR_MOD_NAME))
mod_summary = [(
mod,
len(all_mod_llrs[mod]) if mod in all_mod_llrs else 0,
len(all_can_llrs[mod]) if mod in all_can_llrs else 0,
) for mod in set(all_mod_llrs).union(all_can_llrs)]
LOGGER.info("Data summary:\n\tmod\tmod_N\tcan_N\n" +
"\n".join("\t" + "\t".join(map(str, x)) for x in mod_summary))
output_mods_data(
all_mod_llrs,
all_can_llrs,
args.modified_bases_set,
args.exclude_modified_bases,
args.out_filename,
)