def create_datasets(project):
log = task.logger
config = GlobalConfig(task.conf)
paths = PathsConfig(config)
project_id = project["id"]
log.info("--- [{0}] --------------------------------------------".format(project_id))
project_path = project["path"]
temp_path = project["temp_path"]
datasets_path = paths.project_results_path(project_path)
ensure_path_exists(datasets_path)
sigdb = SigDb(config.sigdb_path)
sigdb.open()
projdb = ProjectDb(project["db"])
gene_sym = projdb.get_gene_symbols()
total_samples = projdb.get_total_affected_samples()
log.info("Exporting variant genes ...")
vf = open_dataset(project_id, project_path, datasets_path, "variant_gene", "w", log)
tsv.write_param(vf, "SAMPLE_TOTAL", total_samples)
tsv.write_line(vf, "VAR_ID", "CHR", "STRAND", "START", "ALLELE",
"GENE_ID", "IMPACT", "IMPACT_CLASS",
"SAMPLE_FREQ", "SAMPLE_PROP",
"CODING_REGION", "PROTEIN_CHANGES", "INTOGEN_DRIVER", "XREFS")
sf = open_dataset(project_id, project_path, datasets_path, "variant-samples", "w", log)
tsv.write_line(sf, "VAR_ID", "CHR", "STRAND", "START", "ALLELE", "SAMPLE")
count = 0
for afg in projdb.affected_genes(join_variant=True, join_samples=True, join_xrefs=True, join_rec=True):
var = afg.var
rec = afg.rec
start, end, ref, alt = var_to_tab(var)
allele = "{0}/{1}".format(ref, alt)
xrefs = [xref for xref in var.xrefs]
if sigdb.exists_variant(var.chr, start):
xrefs += ["I:1"]
xrefs = ",".join(xrefs)
intogen_driver = 1 if sigdb.exists_gene(afg.gene_id) else 0
tsv.write_line(vf, var.id, var.chr, var.strand, start, allele,
afg.gene_id, afg.impact, TransFIC.class_name(afg.impact),
rec.sample_freq, rec.sample_prop,
afg.coding_region, afg.prot_changes, intogen_driver, xrefs, null_value="\N")
for sample in var.samples:
tsv.write_line(sf, var.id, var.chr, var.strand, start, allele, sample.name, null_value="\N")
count += 1
vf.close()
sf.close()
log.info(" {0} variant genes".format(count))
log.info("Exporting consequences ...")
cf = open_dataset(project_id, project_path, datasets_path, "consequence", "w", log)
tsv.write_line(cf, "VAR_ID", "CHR", "STRAND", "START", "ALLELE", "TRANSCRIPT_ID",
"CT", "GENE_ID", "SYMBOL", "UNIPROT_ID", "PROTEIN_ID", "PROTEIN_POS", "AA_CHANGE",
"SIFT_SCORE", "SIFT_TRANSFIC", "SIFT_TRANSFIC_CLASS",
"PPH2_SCORE", "PPH2_TRANSFIC", "PPH2_TRANSFIC_CLASS",
"MA_SCORE", "MA_TRANSFIC", "MA_TRANSFIC_CLASS",
"IMPACT", "IMPACT_CLASS")
count = 0
for csq in projdb.consequences(join_variant=True):
var = csq.var
start, end, ref, alt = var_to_tab(var)
allele = "{0}/{1}".format(ref, alt)
uniprot = protein = protein_pos = aa_change = None
sift_score = sift_tfic = sift_tfic_class = None
pph2_score = pph2_tfic = pph2_tfic_class = None
ma_score = ma_tfic = ma_tfic_class = None
if so.match(csq.ctypes, so.ONCODRIVEFM):
uniprot, protein = csq.uniprot, csq.protein
if so.match(csq.ctypes, so.NON_SYNONYMOUS):
protein_pos, aa_change = csq.protein_pos, csq.aa_change
sift_score, sift_tfic, sift_tfic_class = csq.sift_score, csq.sift_tfic, TransFIC.class_name(csq.sift_tfic_class)
pph2_score, pph2_tfic, pph2_tfic_class = csq.pph2_score, csq.pph2_tfic, TransFIC.class_name(csq.pph2_tfic_class)
ma_score, ma_tfic, ma_tfic_class = csq.ma_score, csq.ma_tfic, TransFIC.class_name(csq.ma_tfic_class)
tsv.write_line(cf, var.id, var.chr, var.strand, start, allele, csq.transcript,
",".join(csq.ctypes), csq.gene, gene_sym.get(csq.gene),
uniprot, protein, protein_pos, aa_change,
sift_score, sift_tfic, sift_tfic_class,
pph2_score, pph2_tfic, pph2_tfic_class,
ma_score, ma_tfic, ma_tfic_class,
csq.impact, TransFIC.class_name(csq.impact), null_value="\N")
count += 1
cf.close()
log.info(" {0} consequences".format(count))
log.info("Exporting genes ...")
gf = open_dataset(project_id, project_path, datasets_path, "gene", "w", log)
tsv.write_param(gf, "SAMPLE_TOTAL", total_samples)
tsv.write_line(gf, "GENE_ID", "FM_PVALUE", "FM_QVALUE", "FM_EXC_CAUSE",
"CLUST_ZSCORE", "CLUST_PVALUE", "CLUST_QVALUE", "CLUST_EXC_CAUSE", "CLUST_COORDS",
"SAMPLE_FREQ", "SAMPLE_PROP", "INTOGEN_DRIVER")
for gene in projdb.genes(join_rec=True):
rec = gene.rec
if rec.sample_freq is None or rec.sample_freq == 0:
continue
intogen_driver = 1 if sigdb.exists_gene(gene.id) else 0
tsv.write_line(gf, gene.id, gene.fm_pvalue, gene.fm_qvalue, gene.fm_exc_cause,
gene.clust_zscore, gene.clust_pvalue, gene.clust_qvalue, gene.clust_exc_cause, gene.clust_coords,
rec.sample_freq or 0, rec.sample_prop or 0,
intogen_driver, null_value="\N")
gf.close()
log.info("Exporting pathways ...")
pf = open_dataset(project_id, project_path, datasets_path, "pathway", "w", log)
tsv.write_param(pf, "SAMPLE_TOTAL", total_samples)
tsv.write_line(pf, "PATHWAY_ID", "GENE_COUNT", "FM_ZSCORE", "FM_PVALUE", "FM_QVALUE",
"SAMPLE_FREQ", "SAMPLE_PROP", "GENE_FREQ", "GENE_TOTAL", "GENE_PROP")
for pathway in projdb.pathways(join_rec=True):
rec = pathway.rec
if rec.sample_freq is None or rec.sample_freq == 0:
continue
tsv.write_line(pf, pathway.id, pathway.gene_count, pathway.fm_zscore, pathway.fm_pvalue, pathway.fm_qvalue,
rec.sample_freq or 0, rec.sample_prop or 0, rec.gene_freq or 0, pathway.gene_count, rec.gene_prop or 0, null_value="\N")
pf.close()
if not config.skip_oncodrivefm:
log.info("Exporting genes per sample functional impact ...")
with open_dataset(project_id, project_path, datasets_path, "gene_sample-fimpact", "w", log) as f:
tsv.write_line(f, "GENE_ID", "SAMPLE",
"SIFT_SCORE", "SIFT_TRANSFIC", "SIFT_TRANSFIC_CLASS",
"PPH2_SCORE", "PPH2_TRANSFIC", "PPH2_TRANSFIC_CLASS",
"MA_SCORE", "MA_TRANSFIC", "MA_TRANSFIC_CLASS")
for fields in projdb.sample_gene_fimpacts():
(gene, sample,
sift_score, sift_tfic, sift_tfic_class,
pph2_score, pph2_tfic, pph2_tfic_class,
ma_score, ma_tfic, ma_tfic_class) = fields
tsv.write_line(f, gene, sample,
sift_score, sift_tfic, TransFIC.class_name(sift_tfic_class),
pph2_score, pph2_tfic, TransFIC.class_name(pph2_tfic_class),
ma_score, ma_tfic, TransFIC.class_name(ma_tfic_class), null_value="\N")
projdb.close()
sigdb.close()
log.info("Saving project configuration ...")
projres = ProjectResults(project)
with open_dataset(project_id, project_path, datasets_path, "project.tsv", "w", log) as f:
names = ["ASSEMBLY", "SAMPLES_TOTAL"]
values = [project["assembly"], total_samples]
names, values = projres.get_annotations_to_save(config.project.annotations, project["annotations"], names=names, values=values)
tsv.write_line(f, *names)
tsv.write_line(f, *values, null_value="\N")
projects_port = task.ports("projects_out")
projects_port.send(project)
def pack_datasets(project):
log = task.logger
config = GlobalConfig(task.conf)
project_id = project["id"]
log.info("--- [{0}] --------------------------------------------".format(project_id))
if not config.results.create_zip:
log.info("Creation of the results compressed file is deactivated. Skipped.")
return
project_path = project["path"]
temp_path = project["temp_path"]
dest_path = os.path.join(project_path, "results.zip")
sigdb = SigDb(config.sigdb_path)
sigdb.open()
projdb = ProjectDb(project["db"])
projres = ProjectResults(project)
gene_sym = projdb.get_gene_symbols()
total_samples = projdb.get_total_affected_samples()
log.info("Compressing files ...")
arc = None
try:
arc = Archive(dest_path, mode="w", fmt="zip")
log.info(" Variant genes ...")
with ArcFile(task, arc, project_id, "variant_genes", "w") as vf:
write_line(vf, "PROJECT_ID", "CHR", "STRAND", "START", "ALLELE",
"GENE_ID", "SYMBOL", "VAR_IMPACT", "VAR_IMPACT_DESC",
"SAMPLE_FREQ", "SAMPLE_TOTAL", "SAMPLE_PROP",
"CODING_REGION", "PROTEIN_CHANGES", "INTOGEN_DRIVER", "XREFS")
for afg in projdb.affected_genes(join_variant=True, join_xrefs=True, join_rec=True):
var = afg.var
rec = afg.rec
start, end, ref, alt = var_to_tab(var)
xrefs = [xref for xref in var.xrefs]
if sigdb.exists_variant(var.chr, start):
xrefs += ["I:1"]
xrefs = ",".join(xrefs)
intogen_driver = 1 if sigdb.exists_gene(afg.gene_id) else 0
write_line(vf, project_id, var.chr, var.strand, start, "{0}/{1}".format(ref, alt),
afg.gene_id, gene_sym.get(afg.gene_id),
afg.impact, TransFIC.class_name(afg.impact),
rec.sample_freq or 0, total_samples, rec.sample_prop or 0,
afg.coding_region, afg.prot_changes, intogen_driver, xrefs)
log.info(" Variant samples ...")
with ArcFile(task, arc, project_id, "variant_samples", "w") as vf:
write_line(vf, "PROJECT_ID", "CHR", "STRAND", "START", "ALLELE", "SAMPLES")
for var in projdb.variants(join_samples=True):
start, end, ref, alt = var_to_tab(var)
write_line(vf, project_id, var.chr, var.strand, start, "{0}/{1}".format(ref, alt),
",".join([s.name for s in var.samples]))
log.info(" Consequences ...")
with ArcFile(task, arc, project_id, "consequences", "w") as cf:
write_line(cf, "PROJECT_ID", "CHR", "STRAND", "START", "ALLELE", "TRANSCRIPT_ID", "CT",
"GENE_ID", "SYMBOL", "UNIPROT_ID", "PROTEIN_ID", "PROTEIN_POS", "AA_CHANGE",
"SIFT_SCORE", "SIFT_TRANSFIC", "SIFT_TRANSFIC_CLASS",
"PPH2_SCORE", "PPH2_TRANSFIC", "PPH2_TRANSFIC_CLASS",
"MA_SCORE", "MA_TRANSFIC", "MA_TRANSFIC_CLASS",
"IMPACT", "IMPACT_CLASS")
for csq in projdb.consequences(join_variant=True):
var = csq.var
start, end, ref, alt = var_to_tab(var)
allele = "{0}/{1}".format(ref, alt)
uniprot = protein = protein_pos = aa_change = None
sift_score = sift_tfic = sift_tfic_class = None
pph2_score = pph2_tfic = pph2_tfic_class = None
ma_score = ma_tfic = ma_tfic_class = None
if so.match(csq.ctypes, so.ONCODRIVEFM):
uniprot, protein = csq.uniprot, csq.protein
if so.match(csq.ctypes, so.NON_SYNONYMOUS):
protein_pos, aa_change = csq.protein_pos, csq.aa_change
sift_score, sift_tfic, sift_tfic_class = csq.sift_score, csq.sift_tfic, TransFIC.class_name(csq.sift_tfic_class)
pph2_score, pph2_tfic, pph2_tfic_class = csq.pph2_score, csq.pph2_tfic, TransFIC.class_name(csq.pph2_tfic_class)
ma_score, ma_tfic, ma_tfic_class = csq.ma_score, csq.ma_tfic, TransFIC.class_name(csq.ma_tfic_class)
write_line(cf, project_id, var.chr, var.strand, start, allele, csq.transcript,
",".join(csq.ctypes), csq.gene, gene_sym.get(csq.gene),
uniprot, protein, protein_pos, aa_change,
sift_score, sift_tfic, sift_tfic_class,
pph2_score, pph2_tfic, pph2_tfic_class,
ma_score, ma_tfic, ma_tfic_class,
csq.impact, TransFIC.class_name(csq.impact))
log.info(" Genes ...")
with ArcFile(task, arc, project_id, "genes", "w") as gf:
write_line(gf, "PROJECT_ID", "GENE_ID", "SYMBOL", "FM_PVALUE", "FM_QVALUE", "FM_EXC_CAUSE",
"SAMPLE_FREQ", "SAMPLE_TOTAL", "SAMPLE_PROP",
"CLUST_ZSCORE", "CLUST_PVALUE", "CLUST_QVALUE", "CLUST_EXC_CAUSE", "CLUST_COORDS",
"INTOGEN_DRIVER", "XREFS")
for gene in projdb.genes(join_xrefs=True, join_rec=True):
if gene.rec.sample_freq is not None and gene.rec.sample_freq > 0:
intogen_driver = 1 if sigdb.exists_gene(gene.id) else 0
write_line(gf, project_id, gene.id, gene.symbol, gene.fm_pvalue, gene.fm_qvalue, gene.fm_exc_cause,
gene.rec.sample_freq, total_samples, gene.rec.sample_prop or 0,
gene.clust_zscore, gene.clust_pvalue, gene.clust_qvalue, gene.clust_exc_cause,
gene.clust_coords, intogen_driver, ",".join(gene.xrefs))
log.info(" Pathways ...")
with ArcFile(task, arc, project_id, "pathways", "w") as pf:
write_line(pf, "PROJECT_ID", "PATHWAY_ID", "GENE_COUNT", "FM_ZSCORE", "FM_PVALUE", "FM_QVALUE",
"SAMPLE_FREQ", "SAMPLE_TOTAL", "SAMPLE_PROP", "GENE_FREQ", "GENE_TOTAL", "GENE_PROP")
for pathway in projdb.pathways(join_rec=True):
if pathway.rec.sample_freq is not None and pathway.rec.sample_freq > 0:
write_line(pf, project_id, pathway.id, pathway.gene_count, pathway.fm_zscore, pathway.fm_pvalue, pathway.fm_qvalue,
pathway.rec.sample_freq or 0, total_samples, pathway.rec.sample_prop or 0,
pathway.rec.gene_freq or 0, pathway.gene_count, pathway.rec.gene_prop or 0)
if not config.skip_oncodrivefm:
log.info(" Genes per sample functional impact ...")
with ArcFile(task, arc, project_id, "fimpact.gitools.tdm", "w") as f:
write_line(f, "SAMPLE", "GENE_ID",
"SIFT_SCORE", "SIFT_TRANSFIC", "SIFT_TRANSFIC_CLASS",
"PPH2_SCORE", "PPH2_TRANSFIC", "PPH2_TRANSFIC_CLASS",
"MA_SCORE", "MA_TRANSFIC", "MA_TRANSFIC_CLASS")
for fields in projdb.sample_gene_fimpacts():
(gene, sample,
sift_score, sift_tfic, sift_tfic_class,
pph2_score, pph2_tfic, pph2_tfic_class,
ma_score, ma_tfic, ma_tfic_class) = fields
write_line(f, sample, gene,
sift_score, sift_tfic, TransFIC.class_name(sift_tfic_class),
pph2_score, pph2_tfic, TransFIC.class_name(pph2_tfic_class),
ma_score, ma_tfic, TransFIC.class_name(ma_tfic_class))
log.info("Saving project configuration ...")
with ArcFile(task, arc, project_id, "project", "w") as f:
names = ["PROJECT_ID", "ASSEMBLY", "SAMPLES_TOTAL"]
values = [project_id, project["assembly"], total_samples]
names, values = projres.get_annotations_to_save(config.project.annotations, project["annotations"], names=names, values=values)
tsv.write_line(f, *names)
tsv.write_line(f, *values, null_value="-")
finally:
if arc is not None:
arc.close()
projdb.close()
sigdb.close()
def drivers():
log = task.logger
config = GlobalConfig(task.conf)
paths = PathsConfig(config)
db_path = paths.results_path("drivers.db")
db = SigDb(db_path)
db.open()
log.info("Variants ...")
path = paths.combination_path("recurrences", "variant_gene-global-all.tsv.gz")
with tsv.open(path, "r") as f:
types = (str, str, int, str)
for fields in tsv.lines(f, types, columns=("CHR", "STRAND", "START", "ALLELE"), header=True):
chr, strand, start, allele = fields[:4]
db.add_variant(chr, start)
log.info("Genes ...")
gene_sites = {}
gene_fm = set()
gene_clust = set()
#SPECIAL_THRESHOLD = ["C18", "C34"]
SPECIAL_THRESHOLD = []
log.info(" OncodriveFM ...")
filename_re = re.compile(r"gene-cancer_site-(.+)\.tsv.gz")
base_path = paths.combination_path("oncodrivefm")
for path in os.listdir(base_path):
m = filename_re.match(path)
if not m:
continue
cancer_site_code = m.group(1)
if cancer_site_code in SPECIAL_THRESHOLD:
threshold = 1e-6
else:
threshold = 0.01
with tsv.open(os.path.join(base_path, path), "r") as f:
params = tsv.params(f)
cancer_site_name = params["group_long_name"]
for fields in tsv.lines(f, (str, float), columns=("ID", "QVALUE"), header=True):
gene, qvalue = fields
if qvalue < threshold:
add_cancer_site(gene_sites, gene, cancer_site_code, cancer_site_name)
gene_fm.add(gene)
log.info(" OncodriveCLUST ...")
filename_re = re.compile(r"cancer_site-(.+)\.tsv.gz")
base_path = paths.combination_path("oncodriveclust")
for path in os.listdir(base_path):
m = filename_re.match(path)
if not m:
continue
cancer_site_code = m.group(1)
with tsv.open(os.path.join(base_path, path), "r") as f:
params = tsv.params(f)
cancer_site_name = params["group_long_name"]
for fields in tsv.lines(f, (str, float), columns=("ID", "QVALUE"), header=True):
gene, qvalue = fields
if qvalue < 0.05:
add_cancer_site(gene_sites, gene, cancer_site_code, cancer_site_name)
gene_clust.add(gene)
log.info(" Updating db ...")
sig_genes = gene_fm | gene_clust
for gene in sig_genes:
db.add_gene(gene, gene in gene_fm, gene in gene_clust)
log.info("Saving driver genes cancer sites dataset ...")
path = paths.results_path("gene-driver_cancer_sites.tsv")
log.debug("> {}".format(path))
with open(path, "w") as f:
tsv.write_param(f, "date", datetime.now())
tsv.write_line(f, "GENE_ID", "FM", "CLUST", "CANCER_SITES_COUNT", "CANCER_SITE_CODES", "CANCER_SITE_NAMES")
for gene, sites in gene_sites.items():
tsv.write_line(f, gene,
1 if gene in gene_fm else 0,
1 if gene in gene_clust else 0,
len(sites),
", ".join(sorted([code for code, name in sites])),
", ".join(sorted([name for code, name in sites])))
db.commit()
db.close()
