def test_rare_2_samp(self):
'''Check that correct sequences are identified as rare when a cut-off
of 2 samples is used.'''
seqtab_in = biom_to_pandas_df(biom.load_table(seqtab_biom))
rare_seqs = id_rare_seqs(seqtab_in, 1, 2)
self.assertSetEqual(set(rare_seqs), set(["2558860574", "2571042244"]))
def test_rare_4_reads(self):
'''Check that correct sequences are identified as rare when a cut-off
of 4 reads is used.'''
seqtab_in = biom_to_pandas_df(biom.load_table(seqtab_biom))
rare_seqs = id_rare_seqs(seqtab_in, 4, 1)
self.assertSetEqual(set(rare_seqs), set(["2558860574", "extra"]))
def test_norm_by_marker_copies(self):
'''Test that expected normalized sequence abundance table generated.'''
seqtab_in = biom_to_pandas_df(biom.load_table(seqtab_biom))
# Get output index labels in same order as expected.
seqtab_in = seqtab_in.reindex(exp_norm_in.index)
test_norm = norm_by_marker_copies(input_seq_counts=seqtab_in,
input_marker_num=marker_predict_in,
norm_filename=None)
# Test whether normalized table matches expected table.
pd.testing.assert_frame_equal(test_norm, exp_norm_in)
def main():
parser = argparse.ArgumentParser(
description="Creates output FASTA for each sample with each ASV repeated for every count in that sample.",
formatter_class=argparse.RawDescriptionHelpFormatter)
parser.add_argument("-f", "--fasta", metavar="FASTA", type=str,
help="Path to full FASTA file.", required=True)
parser.add_argument("-b", "--biom", metavar="BIOM", type=str,
help="Path to BIOM table.", required=True)
parser.add_argument("-o", "--outdir", metavar="PATH", type=str,
help="Name of folder to make for output files.", required=True)
args = parser.parse_args()
in_fasta = read_fasta(args.fasta)
in_table = biom_to_pandas_df(biom.load_table(args.biom))
# If no sequences in file then stop job.
if not in_fasta:
sys.exit("Stopping - no sequences in file.")
make_output_dir(args.outdir)
for sample in in_table.columns:
sample_outfile = args.outdir + "/" + sample + ".fasta"
sample_outfh = open(sample_outfile, 'wt')
for asv in in_table.index.values:
asv_count = in_table.loc[asv, sample]
if asv_count > 0:
for i in range(int(asv_count)):
print(">" + asv + "_" + sample + "_" + str(i), file=sample_outfh)
print(in_fasta[asv], file=sample_outfh)
sample_outfh.close()
def run_metagenome_pipeline(input_biom,
function,
marker,
max_nsti,
out_dir='metagenome_out',
proc=1,
output_normfile=False):
'''Main function to run full metagenome pipeline. Meant to run modular
functions largely listed below. Will return predicted metagenomes
straitifed and unstratified by contributing genomes (i.e. taxa).'''
# Read in input table of sequence abundances and convert to pandas df.
study_seq_counts = biom_to_pandas_df(biom.load_table(input_biom))
# Read in predicted function and marker gene abundances.
pred_function = pd.read_table(function, sep="\t", index_col="sequence")
pred_marker = pd.read_table(marker, sep="\t", index_col="sequence")
pred_function.index = pred_function.index.astype(str)
pred_marker.index = pred_marker.index.astype(str)
# Initialize empty pandas dataframe to contain NSTI values.
nsti_val = pd.DataFrame()
# If NSTI column present then remove all rows with value above specified
# max value. Also, remove NSTI column (in both dataframes).
if "metadata_NSTI" in pred_function.columns:
pred_function = pred_function[pred_function['metadata_NSTI'] <= max_nsti]
nsti_val = pred_function[['metadata_NSTI']]
pred_function.drop('metadata_NSTI', axis=1, inplace=True)
if "metadata_NSTI" in pred_marker.columns:
pred_marker = pred_marker[pred_marker['metadata_NSTI'] <= max_nsti]
nsti_val = pred_marker[['metadata_NSTI']]
pred_marker.drop('metadata_NSTI', axis=1, inplace=True)
# Re-order predicted abundance tables to be in same order as study seqs.
# Also, drop any sequence ids that don't overlap across all dataframes.
study_seq_counts, pred_function, pred_marker = three_df_index_overlap_sort(study_seq_counts,
pred_function,
pred_marker)
# Create output directory if it does not already exist.
make_output_dir(out_dir)
# Create normalized sequence abundance filename if outfile specified.
if output_normfile:
norm_output = path.join(out_dir, "seqtab_norm.tsv")
else:
norm_output = None
# Normalize input study sequence abundances by predicted abundance of
# marker genes and output normalized table if specified.
study_seq_counts = norm_by_marker_copies(input_seq_counts=study_seq_counts,
input_marker_num=pred_marker,
norm_filename=norm_output)
# If NSTI column input then output weighted NSTI values.
if not nsti_val.empty:
weighted_nsti_out = path.join(out_dir, "weighted_nsti.tsv")
weighted_nsti = calc_weighted_nsti(seq_counts=study_seq_counts,
nsti_input=nsti_val,
outfile=weighted_nsti_out)
# Get predicted function counts by sample, stratified by contributing
# genomes and also separately unstratified.
return(funcs_by_sample(input_seq_counts=study_seq_counts,
input_function_num=pred_function,
proc=proc))