def check_missing(args, data, features):
m = None
if args.missing_individuals:
logging.info("Instructed to check for individuals missing from the genotype")
p = Parquet._read(data, ["individual"])
g = Parquet._read(features, ["individual"])
m = [x for x in p["individual"] if x in g["individual"]]
logging.info("Found %d individuals", len(m))
return m
def process(w, s, c, data, data_annotation_, features, features_metadata, x_weights, summary_fields, train, postfix=None, nested_folds=10, use_individuals=None):
gene_id_ = data_annotation_.gene_id if postfix is None else "{}-{}".format(data_annotation_.gene_id, postfix)
logging.log(8, "loading data")
d_ = Parquet._read(data, [data_annotation_.gene_id], specific_individuals=use_individuals)
features_ = Genomics.entries_for_gene_annotation(data_annotation_, args.window, features_metadata)
if x_weights is not None:
x_w = features_[["id"]].merge(x_weights[x_weights.gene_id == data_annotation_.gene_id], on="id")
features_ = features_[features_.id.isin(x_w.id)]
x_w = robjects.FloatVector(x_w.w.values)
else:
x_w = None
if features_.shape[0] == 0:
logging.log(9, "No features available")
return
features_data_ = Parquet._read(features, [x for x in features_.id.values],
specific_individuals=[x for x in d_["individual"]])
logging.log(8, "training")
weights, summary = train(features_data_, features_, d_, data_annotation_, x_w, not args.dont_prune, nested_folds)
if weights.shape[0] == 0:
logging.log(9, "no weights, skipping")
return
logging.log(8, "saving")
weights = weights.assign(gene=data_annotation_.gene_id). \
merge(features_.rename(columns={"id": "feature", "allele_0": "ref_allele", "allele_1": "eff_allele"}), on="feature"). \
rename(columns={"feature": "varID"}). \
assign(gene=gene_id_)
weights = weights[["gene", "rsid", "varID", "ref_allele", "eff_allele", "weight"]]
if args.output_rsids:
weights.loc[weights.rsid == "NA", "rsid"] = weights.loc[weights.rsid == "NA", "varID"]
w.write(weights.to_csv(sep="\t", index=False, header=False, na_rep="NA").encode())
summary = summary. \
assign(gene=gene_id_, genename=data_annotation_.gene_name,
gene_type=data_annotation_.gene_type). \
rename(columns={"n_features": "n_snps_in_window", "n_features_in_model": "n.snps.in.model",
"zscore_pval": "pred.perf.pval", "rho_avg_squared": "pred.perf.R2",
"cv_converged":"nested_cv_converged"})
summary["pred.perf.qval"] = None
summary = summary[summary_fields]
s.write(summary.to_csv(sep="\t", index=False, header=False, na_rep="NA").encode())
var_ids = [x for x in weights.varID.values]
cov = numpy.cov([features_data_[k] for k in var_ids], ddof=1)
ids = [x for x in weights.rsid.values] if args.output_rsids else var_ids
cov = matrices._flatten_matrix_data([(gene_id_, ids, cov)])
for cov_ in cov:
l = "{} {} {} {}\n".format(cov_[0], cov_[1], cov_[2], cov_[3]).encode()
c.write(l)
def by_chromosome(context, chromosome):
vm = context.vmf.read_row_group(chromosome - 1).to_pandas()
if args.frequency_filter:
vm = filter_by_frequency(vm, args.frequency_filter)
g = context.get_genotype_file(chromosome)
regions = context.regions
regions = regions[regions.chr == "chr{}".format(chromosome)]
for i, region in enumerate(regions.itertuples()):
logging.log(9, "Processing region in chr %d: %d/%d", chromosome, i + 1,
regions.shape[0])
vmw = Genomics.entries_for_window(chromosome,
region.start - args.window,
region.stop + args.window, vm)
ids = vmw.id.values
logging.log(9, "%d variants", len(ids))
d = Parquet._read(g, columns=ids, skip_individuals=True)
d = numpy.array([d[x] for x in ids], dtype=numpy.float32)
if context.args.standardise_geno:
cov = numpy.corrcoef(d, ddof=1).astype(numpy.float32, copy=False)
else:
cov = numpy.cov(d).astype(numpy.float32, copy=False)
logging.log(9, "%d rows", cov.shape[0])
context.sink(cov, ids, region)
def run(args):
if os.path.exists(args.output):
logging.info("Output already exists, either delete it or move it")
return
logging.info("Getting parquet genotypes")
file_map = get_file_map(args)
logging.info("Getting genes")
with sqlite3.connect(args.model_db) as connection:
extra = pandas.read_sql("SELECT * FROM EXTRA", connection)
logging.info("Processing")
with gzip.open(args.output, "w") as f:
f.write("GENE RSID1 RSID2 VALUE\n".encode())
with sqlite3.connect(args.model_db) as connection:
for i,t in enumerate(extra.itertuples()):
g_ = t.gene
logging.log(9, "Proccessing %i:%s", i, g_)
w = pandas.read_sql("select * from weights where gene = '{}';".format(g_), connection)
chr_ = w.varID.values[0].split("_")[0].split("chr")[1]
dosage = file_map[int(chr_)]
d = Parquet._read(dosage, columns=w.varID.values, skip_individuals=True)
var_ids = list(d.keys())
rsids = pandas.DataFrame({"varID":var_ids}).merge(w[["varID", "rsid"]], on="varID").rsid
c = numpy.cov([d[x] for x in var_ids])
c = matrices._flatten_matrix_data([(w.gene.values[0], rsids, c)])
for entry in c:
l = "{} {} {} {}\n".format(entry[0], entry[1], entry[2], entry[3])
f.write(l.encode())
logging.info("Finished building covariance.")
def run(args):
if os.path.exists(args.output):
logging.info("Output already exists, either delete it or move it")
return
logging.info("Getting parquet genotypes")
file_map = get_file_map(args)
logging.info("Getting variants")
gene_variants = get_gene_variant_list(args.model_db_folder,
args.model_db_file_pattern)
genes = list(gene_variants.gene.drop_duplicates())
Utilities.ensure_requisite_folders(args.output)
logging.info("Processing")
with gzip.open(args.output, "w") as f:
f.write("GENE RSID1 RSID2 VALUE\n".encode())
for i, g in enumerate(gene_variants.gene.drop_duplicates()):
logging.log(9, "Proccessing %i/%i:%s", i + 1, len(genes), g)
w = gene_variants[gene_variants.gene == g]
chr_ = w.varID.values[0].split("_")[0].split("chr")[1]
if not n_.search(chr_):
logging.log(9, "Unsupported chromosome: %s", chr_)
continue
dosage = file_map[int(chr_)]
d = Parquet._read(dosage,
columns=w.varID.values,
skip_individuals=True)
var_ids = list(d.keys())
if args.output_rsids:
ids = [
x for x in pandas.DataFrame({
"varID": var_ids
}).merge(w[["varID", "rsid"]], on="varID").rsid.values
]
else:
ids = var_ids
c = numpy.cov([d[x] for x in var_ids])
c = matrices._flatten_matrix_data([(w.gene.values[0], ids, c)])
for entry in c:
l = "{} {} {} {}\n".format(entry[0], entry[1], entry[2],
entry[3])
f.write(l.encode())
logging.info("Finished building covariance.")
def _run_dap(region, features, features_metadata, summary_stats,
intermediate_folder, output_folder, options, dap_command):
logging.log(9, "fetching and prepatring data")
os.makedirs(_intermediate_folder(intermediate_folder, region))
s = summary_stats[summary_stats.region_id == region.region_id]
#m = features_metadata[features_metadata.id.isin(s.variant_id)]
x = Parquet._read(features, [x for x in s.variant_id.values])
c = numpy.corrcoef([x[k] for k in s.variant_id.values])
del x
numpy.savetxt(_cor_path(intermediate_folder, region), c)
s[["variant_id", "z"]].rename(columns={
"variant_id": "snp_name_i",
"z": "z_i"
}).to_csv(_stats_path(intermediate_folder, region), index=False, sep="\t")
del s
del c
command = _dap_command(region, intermediate_folder, output_folder, options,
dap_command)
logging.log(9, "running")
script_path = _script_path(intermediate_folder, region)
with open(script_path, "w") as script:
script.write(command)
_o = os.path.join(_intermediate_folder(intermediate_folder, region),
"dap.o")
_e = os.path.join(_intermediate_folder(intermediate_folder, region),
"dap.e")
with open(_o, "w") as o:
with open(_e, "w") as e:
call(["bash", script_path], stderr=e, stdout=o)
shutil.move(_o, _output(output_folder, region))
logging.log(9, "executed dap")
def run(args):
Utilities.maybe_create_folder(args.intermediate_folder)
Utilities.ensure_requisite_folders(args.output_prefix)
logging.info("Opening data")
p_ = re.compile(args.data_name_pattern)
f = [x for x in sorted(os.listdir(args.data_folder)) if p_.search(x)]
tissue_names = [p_.search(x).group(1) for x in f]
data = []
for i in range(0, len(tissue_names)):
logging.info("Loading %s", tissue_names[i])
data.append((tissue_names[i],
pq.ParquetFile(os.path.join(args.data_folder, f[i]))))
data = collections.OrderedDict(data)
available_data = {
x
for p in data.values() for x in p.metadata.schema.names
}
logging.info("Preparing output")
WEIGHTS_FIELDS = [
"gene", "rsid", "varID", "ref_allele", "eff_allele", "weight"
]
SUMMARY_FIELDS = [
"gene", "genename", "gene_type", "alpha", "n_snps_in_window",
"n.snps.in.model", "rho_avg", "pred.perf.R2", "pred.perf.pval"
]
Utilities.ensure_requisite_folders(args.output_prefix)
if args.skip_regression:
weights, summaries, covariances = None, None, None
else:
weights, summaries, covariances = setup_output(args.output_prefix,
tissue_names,
WEIGHTS_FIELDS,
SUMMARY_FIELDS)
logging.info("Loading data annotation")
data_annotation = StudyUtilities._load_gene_annotation(
args.data_annotation)
data_annotation = data_annotation[data_annotation.gene_id.isin(
available_data)]
if args.chromosome or (args.sub_batches and args.sub_batch):
data_annotation = StudyUtilities._filter_gene_annotation(
data_annotation, args.chromosome, args.sub_batches, args.sub_batch)
logging.info("Kept %i entries", data_annotation.shape[0])
logging.info("Opening features annotation")
if not args.chromosome:
features_metadata = pq.read_table(args.features_annotation).to_pandas()
else:
features_metadata = pq.ParquetFile(
args.features_annotation).read_row_group(args.chromosome -
1).to_pandas()
if args.chromosome and args.sub_batches:
logging.info("Trimming variants")
features_metadata = StudyUtilities.trim_variant_metadata_on_gene_annotation(
features_metadata, data_annotation, args.window)
if args.rsid_whitelist:
logging.info("Filtering features annotation")
whitelist = TextFileTools.load_list(args.rsid_whitelist)
whitelist = set(whitelist)
features_metadata = features_metadata[features_metadata.rsid.isin(
whitelist)]
logging.info("Opening features")
features = pq.ParquetFile(args.features)
logging.info("Setting R seed")
seed = numpy.random.randint(1e8)
if args.run_tag:
d = pandas.DataFrame({
"run": [args.run_tag],
"cv_seed": [seed]
})[["run", "cv_seed"]]
for t in tissue_names:
Utilities.save_dataframe(
d, "{}_{}_runs.txt.gz".format(args.output_prefix, t))
failed_run = False
try:
for i, data_annotation_ in enumerate(data_annotation.itertuples()):
logging.log(9, "processing %i/%i:%s", i + 1,
data_annotation.shape[0], data_annotation_.gene_id)
logging.log(8, "loading data")
d_ = {}
for k, v in data.items():
d_[k] = Parquet._read(v, [data_annotation_.gene_id],
to_pandas=True)
features_ = Genomics.entries_for_gene_annotation(
data_annotation_, args.window, features_metadata)
if features_.shape[0] == 0:
logging.log(9, "No features available")
continue
features_data_ = Parquet._read(features,
[x for x in features_.id.values],
to_pandas=True)
features_data_["id"] = range(1, features_data_.shape[0] + 1)
features_data_ = features_data_[["individual", "id"] +
[x for x in features_.id.values]]
logging.log(8, "training")
prepare_ctimp(args.script_path, seed, args.intermediate_folder,
data_annotation_, features_, features_data_, d_)
del (features_data_)
del (d_)
if args.skip_regression:
continue
subprocess.call([
"bash",
_execution_script(args.intermediate_folder,
data_annotation_.gene_id)
])
w = pandas.read_table(_weights(args.intermediate_folder,
data_annotation_.gene_id),
sep="\s+")
s = pandas.read_table(_summary(args.intermediate_folder,
data_annotation_.gene_id),
sep="\s+")
for e_, entry in enumerate(s.itertuples()):
entry_weights = w[["SNP", "REF.0.", "ALT.1.",
entry.tissue]].rename(
columns={
"SNP": "varID",
"REF.0.": "ref_allele",
"ALT.1.": "eff_allele",
entry.tissue: "weight"
})
entry_weights = entry_weights[entry_weights.weight != 0]
entry_weights = entry_weights.assign(
gene=data_annotation_.gene_id)
entry_weights = entry_weights.merge(features_,
left_on="varID",
right_on="id",
how="left")
entry_weights = entry_weights[WEIGHTS_FIELDS]
if args.output_rsids:
entry_weights.loc[entry_weights.rsid == "NA",
"rsid"] = entry_weights.loc[
entry_weights.rsid == "NA", "varID"]
weights[entry.tissue].write(
entry_weights.to_csv(sep="\t",
index=False,
header=False,
na_rep="NA").encode())
entry_summary = s[s.tissue == entry.tissue].rename(
columns={
"zscore_pval": "pred.perf.pval",
"rho_avg_squared": "pred.perf.R2"
})
entry_summary = entry_summary.assign(
gene=data_annotation_.gene_id,
alpha=0.5,
genename=data_annotation_.gene_name,
gene_type=data_annotation_.gene_type,
n_snps_in_window=features_.shape[0])
entry_summary["n.snps.in.model"] = entry_weights.shape[0]
#must repeat strings beause of weird pandas indexing issue
entry_summary = entry_summary.drop(
["R2", "n", "tissue"], axis=1)[[
"gene", "genename", "gene_type", "alpha",
"n_snps_in_window", "n.snps.in.model", "rho_avg",
"pred.perf.R2", "pred.perf.pval"
]]
summaries[entry.tissue].write(
entry_summary.to_csv(sep="\t",
index=False,
header=False,
na_rep="NA").encode())
features_data_ = Parquet._read(
features, [x for x in entry_weights.varID.values],
to_pandas=True)
var_ids = [x for x in entry_weights.varID.values]
cov = numpy.cov([features_data_[k] for k in var_ids], ddof=1)
ids = [x for x in entry_weights.rsid.values
] if args.output_rsids else var_ids
cov = matrices._flatten_matrix_data([(data_annotation_.gene_id,
ids, cov)])
for cov_ in cov:
l = "{} {} {} {}\n".format(cov_[0], cov_[1], cov_[2],
cov_[3]).encode()
covariances[entry.tissue].write(l)
if not args.keep_intermediate_folder:
logging.info("Cleaning up")
shutil.rmtree(
_intermediate_folder(args.intermediate_folder,
data_annotation_.gene_id))
if args.MAX_M and i >= args.MAX_M:
logging.info("Early abort")
break
except Exception as e:
logging.info("Exception running model training:\n%s",
traceback.format_exc())
failed_run = True
finally:
pass
# if not args.keep_intermediate_folder:
# shutil.rmtree(args.intermediate_folder)
if not args.skip_regression:
set_down(weights, summaries, covariances, tissue_names, failed_run)
logging.info("Finished")
def run(args):
if os.path.exists(args.output):
logging.info("Output already exists, either delete it or move it")
return
logging.info("Getting parquet genotypes")
file_map = get_file_map(args)
logging.info("Getting genes")
with sqlite3.connect(args.model_db) as connection:
# Pay heed to the order. This avoids arbitrariness in sqlite3 loading of results.
extra = pandas.read_sql("SELECT * FROM EXTRA order by gene",
connection)
extra = extra[extra["n.snps.in.model"] > 0]
individuals = TextFileTools.load_list(
args.individuals) if args.individuals else None
logging.info("Processing")
Utilities.ensure_requisite_folders(args.output)
with gzip.open(args.output, "w") as f:
f.write("GENE RSID1 RSID2 VALUE\n".encode())
with sqlite3.connect(args.model_db) as connection:
for i, t in enumerate(extra.itertuples()):
g_ = t.gene
logging.log(9, "Proccessing %i/%i:%s", i + 1, extra.shape[0],
g_)
w = pandas.read_sql(
"select * from weights where gene = '{}';".format(g_),
connection)
chr_ = w.varID.values[0].split("_")[0].split("chr")[1]
if not n_.search(chr_):
logging.log(9, "Unsupported chromosome: %s", chr_)
continue
dosage = file_map[int(chr_)]
if individuals:
d = Parquet._read(dosage,
columns=w.varID.values,
specific_individuals=individuals)
del d["individual"]
else:
d = Parquet._read(dosage,
columns=w.varID.values,
skip_individuals=True)
var_ids = list(d.keys())
if len(var_ids) == 0:
if len(w.varID.values) == 1:
logging.log(
9, "workaround for single missing genotype at %s",
g_)
d = {w.varID.values[0]: [0, 1]}
else:
logging.log(9,
"No genotype available for %s, skipping",
g_)
next
if args.output_rsids:
ids = [
x for x in pandas.DataFrame({
"varID": var_ids
}).merge(w[["varID", "rsid"]], on="varID").rsid.values
]
else:
ids = var_ids
c = numpy.cov([d[x] for x in var_ids])
c = matrices._flatten_matrix_data([(w.gene.values[0], ids, c)])
for entry in c:
l = "{} {} {} {}\n".format(entry[0], entry[1], entry[2],
entry[3])
f.write(l.encode())
logging.info("Finished building covariance.")
def run(args):
if os.path.exists(args.output):
logging.info("Output already exists, either delete it or move it")
return
logging.info("Loading group")
groups = pandas.read_table(args.group)
groups = groups.assign(chromosome = groups.gtex_intron_id.str.split(":").str.get(0))
groups = groups.assign(position=groups.gtex_intron_id.str.split(":").str.get(1))
groups = Genomics.sort(groups)
logging.info("Getting parquet genotypes")
file_map = get_file_map(args)
logging.info("Getting genes")
with sqlite3.connect(args.model_db_group_key) as connection:
# Pay heed to the order. This avoids arbitrariness in sqlite3 loading of results.
extra = pandas.read_sql("SELECT * FROM EXTRA order by gene", connection)
extra = extra[extra["n.snps.in.model"] > 0]
individuals = TextFileTools.load_list(args.individuals) if args.individuals else None
logging.info("Processing")
Utilities.ensure_requisite_folders(args.output)
genes_ = groups[["chromosome", "position", "gene_id"]].drop_duplicates()
with gzip.open(args.output, "w") as f:
f.write("GENE RSID1 RSID2 VALUE\n".encode())
with sqlite3.connect(args.model_db_group_key) as db_group_key:
with sqlite3.connect(args.model_db_group_values) as db_group_values:
for i,t_ in enumerate(genes_.itertuples()):
g_ = t_.gene_id
chr_ = t_.chromosome.split("chr")[1]
logging.log(8, "Proccessing %i/%i:%s", i+1, len(genes_), g_)
if not n_.search(chr_):
logging.log(9, "Unsupported chromosome: %s", chr_)
continue
dosage = file_map[int(chr_)]
group = groups[groups.gene_id == g_]
wg=[]
for value in group.intron_id:
wk = pandas.read_sql("select * from weights where gene = '{}';".format(value), db_group_values)
if wk.shape[0] == 0:
continue
wg.append(wk)
if len(wg) > 0:
wg = pandas.concat(wg)
w = pandas.concat([wk, wg])[["varID", "rsid"]].drop_duplicates()
else:
w = wk[["varID", "rsid"]].drop_duplicates()
if w.shape[0] == 0:
logging.log(8, "No data, skipping")
continue
if individuals:
d = Parquet._read(dosage, columns=w.varID.values, specific_individuals=individuals)
del d["individual"]
else:
d = Parquet._read(dosage, columns=w.varID.values, skip_individuals=True)
var_ids = list(d.keys())
if len(var_ids) == 0:
if len(w.varID.values) == 1:
logging.log(9, "workaround for single missing genotype at %s", g_)
d = {w.varID.values[0]:[0,1]}
else:
logging.log(9, "No genotype available for %s, skipping",g_)
next
if args.output_rsids:
ids = [x for x in pandas.DataFrame({"varID": var_ids}).merge(w[["varID", "rsid"]], on="varID").rsid.values]
else:
ids = var_ids
c = numpy.cov([d[x] for x in var_ids])
c = matrices._flatten_matrix_data([(g_, ids, c)])
for entry in c:
l = "{} {} {} {}\n".format(entry[0], entry[1], entry[2], entry[3])
f.write(l.encode())
logging.info("Finished building covariance.")