Mercurial > repos > iuc > vsnp_determine_ref_from_data

annotate vsnp_add_zero_coverage.py @ 7:57bd5b859e86 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
author iuc
date Fri, 10 Jun 2022 06:10:23 +0000
parents b03e88e7bb1d
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
1 #!/usr/bin/env python
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
2
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
3 import argparse
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
4 import os
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
5 import re
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
6 import shutil
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
7
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
8 import pandas
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
9 import pysam
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
10 from Bio import SeqIO
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
11
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
12
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
13 def get_sample_name(file_path):
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
14 base_file_name = os.path.basename(file_path)
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
15 if base_file_name.find(".") > 0:
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
16 # Eliminate the extension.
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
17 return os.path.splitext(base_file_name)[0]
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
18 return base_file_name
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
19
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
20
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
21 def get_coverage_df(bam_file):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
22 # Create a coverage dictionary.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
23 coverage_dict = {}
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
24 coverage_list = pysam.depth(bam_file, split_lines=True)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
25 for line in coverage_list:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
26 chrom, position, depth = line.split('\t')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
27 coverage_dict["%s-%s" % (chrom, position)] = depth
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
28 # Convert it to a data frame.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
29 coverage_df = pandas.DataFrame.from_dict(coverage_dict, orient='index', columns=["depth"])
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
30 return coverage_df
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
31
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
32
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
33 def get_zero_df(reference):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
34 # Create a zero coverage dictionary.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
35 zero_dict = {}
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
36 reference_length = 0
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
37 for record in SeqIO.parse(reference, "fasta"):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
38 chrom = record.id
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
39 total_len = len(record.seq)
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
40 reference_length = reference_length + len(record.seq)
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
41 for pos in list(range(1, total_len + 1)):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
42 zero_dict["%s-%s" % (str(chrom), str(pos))] = 0
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
43 # Convert it to a data frame with depth_x
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
44 # and depth_y columns - index is NaN.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
45 zero_df = pandas.DataFrame.from_dict(zero_dict, orient='index', columns=["depth"])
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
46 return zero_df, reference_length
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
47
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
48
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
49 def output_zc_vcf_file(base_file_name, vcf_file, zero_df, total_zero_coverage, output_vcf):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
50 column_names = ["CHROM", "POS", "ID", "REF", "ALT", "QUAL", "FILTER", "INFO", "FORMAT", "Sample"]
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
51 vcf_df = pandas.read_csv(vcf_file, sep='\t', header=None, names=column_names, comment='#')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
52 good_snp_count = len(vcf_df[(vcf_df['ALT'].str.len() == 1) & (vcf_df['REF'].str.len() == 1) & (vcf_df['QUAL'] > 150)])
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
53 if total_zero_coverage > 0:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
54 header_file = "%s_header.csv" % base_file_name
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
55 with open(header_file, 'w') as outfile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
56 with open(vcf_file) as infile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
57 for line in infile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
58 if re.search('^#', line):
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
59 outfile.write("%s" % line)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
60 vcf_df_snp = vcf_df[vcf_df['REF'].str.len() == 1]
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
61 vcf_df_snp = vcf_df_snp[vcf_df_snp['ALT'].str.len() == 1]
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
62 vcf_df_snp['ABS_VALUE'] = vcf_df_snp['CHROM'].map(str) + "-" + vcf_df_snp['POS'].map(str)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
63 vcf_df_snp = vcf_df_snp.set_index('ABS_VALUE')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
64 cat_df = pandas.concat([vcf_df_snp, zero_df], axis=1, sort=False)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
65 cat_df = cat_df.drop(columns=['CHROM', 'POS', 'depth'])
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
66 cat_df[['ID', 'ALT', 'QUAL', 'FILTER', 'INFO']] = cat_df[['ID', 'ALT', 'QUAL', 'FILTER', 'INFO']].fillna('.')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
67 cat_df['REF'] = cat_df['REF'].fillna('N')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
68 cat_df['FORMAT'] = cat_df['FORMAT'].fillna('GT')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
69 cat_df['Sample'] = cat_df['Sample'].fillna('./.')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
70 cat_df['temp'] = cat_df.index.str.rsplit('-', n=1)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
71 cat_df[['CHROM', 'POS']] = pandas.DataFrame(cat_df.temp.values.tolist(), index=cat_df.index)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
72 cat_df = cat_df[['CHROM', 'POS', 'ID', 'REF', 'ALT', 'QUAL', 'FILTER', 'INFO', 'FORMAT', 'Sample']]
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
73 cat_df['POS'] = cat_df['POS'].astype(int)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
74 cat_df = cat_df.sort_values(['CHROM', 'POS'])
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
75 body_file = "%s_body.csv" % base_file_name
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
76 cat_df.to_csv(body_file, sep='\t', header=False, index=False)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
77 with open(output_vcf, "w") as outfile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
78 for cf in [header_file, body_file]:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
79 with open(cf, "r") as infile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
80 for line in infile:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
81 outfile.write("%s" % line)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
82 else:
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
83 shutil.move(vcf_file, output_vcf)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
84 return good_snp_count
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
85
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
86
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
87 def output_metrics_file(base_file_name, average_coverage, genome_coverage, good_snp_count, dbkey, reference,
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
88 reference_length, total_zero_coverage, percent_ref_with_zero_coverage, output_metrics):
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
89 columns = ["BAM File", "Reference", "reference Length", "Genome with Coverage", "Average Depth",
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
90 "No Coverage Bases", "Percent Ref with Zero Coverage", "Quality SNPs"]
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
91 values = [base_file_name, dbkey, str(reference_length), genome_coverage, average_coverage,
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
92 str(total_zero_coverage), percent_ref_with_zero_coverage, str(good_snp_count)]
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
93 with open(output_metrics, "w") as fh:
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
94 fh.write("# %s\n" % "\t".join(columns))
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
95 fh.write("%s\n" % "\t".join(values))
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
96
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
97
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
98 def output_files(vcf_file, total_zero_coverage, percent_ref_with_zero_coverage, zero_df, output_vcf,
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
99 average_coverage, genome_coverage, output_metrics, reference, reference_length, dbkey):
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
100 base_file_name = get_sample_name(vcf_file)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
101 good_snp_count = output_zc_vcf_file(base_file_name, vcf_file, zero_df, total_zero_coverage, output_vcf)
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
102 output_metrics_file(base_file_name, average_coverage, genome_coverage, good_snp_count, dbkey, reference,
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
103 reference_length, total_zero_coverage, percent_ref_with_zero_coverage, output_metrics)
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
104
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
105
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
106 def get_coverage_and_snp_count(bam_file, vcf_file, dbkey, reference, output_metrics, output_vcf):
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
107 coverage_df = get_coverage_df(bam_file)
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
108 zero_df, reference_length = get_zero_df(reference)
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
109 coverage_df = zero_df.merge(coverage_df, left_index=True, right_index=True, how='outer')
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
110 # depth_x "0" column no longer needed.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
111 coverage_df = coverage_df.drop(columns=['depth_x'])
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
112 coverage_df = coverage_df.rename(columns={'depth_y': 'depth'})
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
113 # Covert the NaN to 0 coverage and get some metrics.
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
114 coverage_df = coverage_df.fillna(0)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
115 coverage_df['depth'] = coverage_df['depth'].apply(int)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
116 total_length = len(coverage_df)
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
117 average_coverage = "{:.2f}".format(coverage_df['depth'].mean())
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
118 zero_df = coverage_df[coverage_df['depth'] == 0]
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
119 total_zero_coverage = len(zero_df)
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
120 percent_ref_with_zero_coverage = "{:.6%}".format(total_zero_coverage / reference_length * 100)
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
121 total_coverage = total_length - total_zero_coverage
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
122 genome_coverage = "{:.2%}".format(total_coverage / total_length)
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
123 # Output a zero-coverage vcf fil and the metrics file.
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
124 output_files(vcf_file, total_zero_coverage, percent_ref_with_zero_coverage, zero_df, output_vcf,
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
125 average_coverage, genome_coverage, output_metrics, reference, reference_length, dbkey)
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
126
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
127
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
128 if __name__ == '__main__':
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
129 parser = argparse.ArgumentParser()
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
130
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
131 parser.add_argument('--bam_input', action='store', dest='bam_input', help='bam input file')
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
132 parser.add_argument('--dbkey', action='store', dest='dbkey', help='bam input dbkey')
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
133 parser.add_argument('--output_metrics', action='store', dest='output_metrics', required=False, default=None, help='Output metrics text file')
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
134 parser.add_argument('--output_vcf', action='store', dest='output_vcf', required=False, default=None, help='Output VCF file')
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
135 parser.add_argument('--reference', action='store', dest='reference', help='Reference dataset')
1
b03e88e7bb1d "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 2e312886647244b416c64eca91e1a61dd1be939b"
iuc
parents: 0
diff changeset
136 parser.add_argument('--vcf_input', action='store', dest='vcf_input', help='vcf input file')
0
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
137
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
138 args = parser.parse_args()
12f2b14549f6 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit 524a39e08f2bea8b8754284df606ff8dd27ed24b"
iuc
parents:
diff changeset
139
7
57bd5b859e86 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/vsnp commit c38fd63f7980c70390d104a73ba4c72b266444c3
iuc
parents: 1
diff changeset
140 get_coverage_and_snp_count(args.bam_input, args.vcf_input, args.dbkey, args.reference, args.output_metrics, args.output_vcf)