def download(args):
"""Downloads fasta or taxonomy dump files
If args.db == 'taxonomy', download taxonomy dump files from ncbi and initialize the ete3 sqlite database
If args.db == 'idmap', download the seqid->taxid mapfile from ncbi
Otherwise download the protein fastafile corresponding to args.db (uniref50, uniref90, uniref100 or nr)
"""
if args.db == "taxonomy":
prepare.download_ncbi_taxonomy(args.taxdir, args.force)
prepare.init_sqlite_taxdb(args.taxdir, args.sqlitedb, args.force)
elif args.db == "idmap":
prepare.download_nr_idmap(args.dldir, args.tmpdir, args.force)
else:
prepare.download_fasta(args.dldir, args.db, args.tmpdir, args.force,
args.skip_check, args.skip_idmap)
def make_lineage_df(taxids, taxdir, dbname, ranks, cpus=1):
"""
Creates a lineage dataframe with full taxonomic information for a list of taxids.
Example:
taxid species phylum genus genus.name phylum.name species.name
859655 305 1224 48736 Ralstonia Proteobacteria Ralstonia solanacearum
387344 1580 1239 1578 Lactobacillus Firmicutes Lactobacillus brevis
358681 1393 1239 55080 Brevibacillus Firmicutes Brevibacillus brevis
Parameters
----------
taxids: list
List of taxonomic ids to obtain information for
taxdir: str
Path to directory holding taxonomic info
dbname: str
Name of ete3 sqlite database within taxdir
ranks: list
Ranks to store information for
cpus: int
Number of cpus to use
Returns
-------
lineage_df: pandas.DataFrame
Data Frame with full taxonomic info
"""
# Read the taxonomy db
ncbi_taxa = init_sqlite_taxdb(taxdir, dbname)
lineages = ncbi_taxa.get_lineage_translator(taxids)
# Store potential missing taxids and warn user
missing_taxids = set([int(x) for x in taxids]).difference(lineages.keys())
# Get possible translations for taxids that have been changed
_, translate_dict = ncbi_taxa._translate_merged(list(set(taxids).difference(lineages.keys())))
rename = {y: x for x, y in translate_dict.items()}
# Update lineages with missing taxids
lineages.update(ncbi_taxa.get_lineage_translator(translate_dict.values()))
items = [[taxid, ranks, taxdir, dbname, lineages[taxid]] for taxid in list(lineages.keys())]
with Pool(processes=cpus) as pool:
res = list(
tqdm.tqdm(pool.imap(process_lineages, items), desc="Making lineages", total=len(items),
unit=" taxids", ncols=100))
lineage_df = pd.concat(res, sort=False)
lineage_df.rename(index=rename, inplace=True)
lineage_df.rename(index=lambda x: int(x), inplace=True)
for rank in ranks:
lineage_df[rank] = pd.to_numeric(lineage_df[rank])
name_dict = make_name_dict(lineage_df, ranks)
if len(missing_taxids) > 0:
sys.stderr.write("#WARNING: Missing taxids found:\n")
sys.stderr.write("#{}\n".format(",".join([str(x) for x in missing_taxids])))
sys.stderr.write("#To fix this, you can try to update the taxonomy database using\n")
sys.stderr.write("#tango download taxonomy --force\n")
return lineage_df.loc[:,lineage_df.dtypes==int], name_dict
def process_lineages(items):
"""
Looks up lineage information from taxids.
The lineage object is a list of taxonomic ids corresponding to the full lineage of a single taxid.
"""
taxid, ranks, taxdir, dbname, lineage = items
# Read the taxonomy db
ncbi_taxa = init_sqlite_taxdb(taxdir, dbname)
# Get ranks for each taxid in the lineage
lineage_ranks = ncbi_taxa.get_rank(lineage)
x = pd.DataFrame(lineage_ranks, index=["rank"]).T
x = x.loc[x["rank"].isin(ranks)].reset_index().T
x.columns = x.loc["rank"]
x.drop("rank", inplace=True)
x.index = [taxid]
# Add taxids for lower ranks in the hierarchy
x = propagate_lower(x, taxid, ranks)
# Add names for taxids
x = add_names(x, taxid, ncbi_taxa)
return x