def _main(args):
logging.init_logger(log_fn=args.log_filename)
if args.ground_truth_cov_min < 1:
LOGGER.warning(
"--ground-truth-cov-min must be 1 or greater. " + "Setting to 1."
)
args.ground_truth_cov_min = 1
if args.nanopore_cov_min < 1:
LOGGER.warning(
"--nanopore-cov-min must be 1 or greater. " + "Setting to 1."
)
args.nanopore_cov_min = 1
LOGGER.info("Checking for consistent contig names")
mod_db_fn = mh.get_megalodon_fn(args.megalodon_results_dir, mh.PR_MOD_NAME)
check_matching_attrs(
args.ground_truth_bed, args.strand_offset, mod_db_fn, args.mod_bases
)
LOGGER.info("Processing batches")
process_all_batches(
args.processes,
args.batch_size,
args.ground_truth_bed,
args.out_low_coverage_sites,
args.out_per_site_mod_thresholds,
mod_db_fn,
args.strand_offset,
args.ground_truth_cov_min,
args.nanopore_cov_min,
args.mod_bases,
args.valid_sites,
)
def _main(args):
logging.init_logger()
pdf_fp = PdfPages(args.out_pdf)
out_fp = (sys.stdout if args.out_filename is None else open(
args.out_filename, 'w'))
(samp1_cov, samp1_mod_cov, samp1_all_cov, samp2_cov, samp2_mod_cov,
samp2_all_cov, valid_pos) = parse_inputs(args.sample1_bed_methyl_files,
args.sample2_bed_methyl_files,
args.strand_offset,
args.sample_names,
args.valid_positions, out_fp)
(samp1_mod_pct, samp2_mod_pct, samp1_valid_cov,
samp2_valid_cov) = compute_filt_mod_pct(samp1_cov, samp1_mod_cov,
samp2_cov, samp2_mod_cov,
valid_pos,
args.coverage_threshold,
args.sample_names, out_fp)
plot_hm(samp1_mod_pct, samp2_mod_pct, args.heatmap_num_bins,
args.sample_names, out_fp, pdf_fp)
plot_cov(samp1_all_cov, samp2_all_cov, samp1_valid_cov, samp2_valid_cov,
args.sample_names, pdf_fp)
pdf_fp.close()
if out_fp is not sys.stdout:
out_fp.close()
def _main(args):
logging.init_logger(log_fn=args.log_filename, quiet=args.quiet)
# add required attributes for loading guppy, but not valid options for
# this script.
args.do_not_use_guppy_server = False
args.output_directory = args.guppy_logs_output_directory
try:
mh.mkdir(args.output_directory, False)
except mh.MegaError:
LOGGER.warning(
"Guppy logs output directory exists. Potentially overwriting "
+ "guppy logs."
)
args = add_trim_guppy_none(args)
args.outputs = [mh.PR_MOD_NAME]
# make edge_buffer >= context_bases to simplify processing
if args.edge_buffer < args.mod_context_bases:
LOGGER.warning(
"[--edge-buffer] less than [--mod-context-bases]. Setting "
+ "[--edge-buffer] to value from [--mod-context-bases]"
)
args.edge_buffer = args.mod_context_bases
LOGGER.info("Loading model.")
backend_params = backends.parse_backend_params(args)
with backends.ModelInfo(backend_params, args.processes) as model_info:
check_map_sig_alphabet(model_info, args.mapped_signal_file)
motifs = parse_motifs(args.motif, model_info, args.modified_bases_set)
can_labs, mod_labs = extract_label_conversions(model_info)
can_post_indices = model_info.can_indices.astype(np.uintp)
all_mod_llrs, all_can_llrs = compute_diff_scores(
args.mapped_signal_file,
model_info,
args.mod_context_bases,
args.edge_buffer,
args.num_reads,
motifs,
can_labs,
mod_labs,
can_post_indices,
)
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.out_filename)
def _main(args):
logging.init_logger()
LOGGER.info("Opening VCF files.")
source_vars = pysam.VariantFile(args.diploid_called_variants)
h1_vars = pysam.VariantFile(args.haplotype1_variants)
h2_vars = pysam.VariantFile(args.haplotype2_variants)
try:
contigs0 = list(source_vars.header.contigs.keys())
source_vars.fetch(contigs0[0])
h1_vars.fetch(next(iter(h1_vars.header.contigs.keys())))
h2_vars.fetch(next(iter(h2_vars.header.contigs.keys())))
except ValueError:
raise mh.MegaError(
"Variant files must be indexed. Use bgzip and tabix.")
LOGGER.info("Processing variants.")
out_vars = open(args.out_vcf, "w")
out_vars.write(
HEADER.format("\n".join(
(CONTIG_HEADER_LINE.format(ctg.name, ctg.length)
for ctg in source_vars.header.contigs.values()))))
bar = tqdm(
total=len(contigs0),
smoothing=0,
unit=" contigs",
dynamic_ncols=True,
desc="Variant Processing",
mininterval=0,
)
for contig in contigs0:
for curr_s_rec, curr_h1_rec, curr_h2_rec in tqdm(
iter_contig_vars(
get_contig_iter(source_vars, contig),
get_contig_iter(h1_vars, contig),
get_contig_iter(h2_vars, contig),
contig,
bar,
args.force_invalid_variant_processing,
),
smoothing=0,
unit=" variants",
dynamic_ncols=True,
leave=False,
desc="{} Variants".format(contig),
):
write_var(curr_s_rec, curr_h1_rec, curr_h2_rec, out_vars, contig)
bar.update(1)
out_vars.close()
variants.index_variants(args.out_vcf)
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,
)
def _main(args):
logging.init_logger()
mh.prep_out_fn(args.out_filename, args.overwrite)
LOGGER.info("Processing {}".format(
args.modified_base_calibration_files[-1]))
calib_data = np.load(args.modified_base_calibration_files[-1])
stratify_type = str(calib_data[calibration.MOD_STRAT_TYPE_TXT])
num_calib_vals = np.int(calib_data[calibration.SMOOTH_NVALS_TXT])
mod_calibs = {}
for mod_base in calib_data[calibration.MOD_BASES_TXT]:
LOGGER.info("\tUsing {} calibration".format(mod_base))
mod_calibs[mod_base] = (
calib_data[mod_base + calibration.LLR_RANGE_SUFFIX].copy(),
calib_data[mod_base + calibration.CALIB_TABLE_SUFFIX].copy(),
)
for mod_calib_fn in args.modified_base_calibration_files[-2::-1]:
LOGGER.info("Processing {}".format(mod_calib_fn))
calib_data = np.load(mod_calib_fn)
assert stratify_type == str(calib_data[calibration.MOD_STRAT_TYPE_TXT])
assert num_calib_vals == np.int(
calib_data[calibration.SMOOTH_NVALS_TXT])
for mod_base in calib_data[calibration.MOD_BASES_TXT]:
# overwrite calibration data with files passed earlier
if mod_base in mod_calibs:
LOGGER.info("\tOverwriting {} calibration".format(mod_base))
else:
LOGGER.info("\tUsing {} calibration".format(mod_base))
mod_calibs[mod_base] = (
calib_data[mod_base + calibration.LLR_RANGE_SUFFIX].copy(),
calib_data[mod_base + calibration.CALIB_TABLE_SUFFIX].copy(),
)
save_kwargs = {}
for mod_base, (mod_llr_range, mod_calib) in mod_calibs.items():
save_kwargs[mod_base + calibration.LLR_RANGE_SUFFIX] = mod_llr_range
save_kwargs[mod_base + calibration.CALIB_TABLE_SUFFIX] = mod_calib
# save calibration table for reading into mod calibration table
LOGGER.info("Saving calibrations to file.")
mod_bases = list(mod_calibs.keys())
np.savez(
args.out_filename,
stratify_type=stratify_type,
smooth_nvals=num_calib_vals,
mod_bases=mod_bases,
**save_kwargs,
)