Mercurial > repos > mvdbeek > size_distribution

comparison size_distributions.py @ 4:f1eeaf42144b 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/bardin-lab/smallRNA_tools commit c8e0a703fcdff580ba0a0c5806a37c088c03ab7b-dirty
author mvdbeek
date Mon, 20 Aug 2018 14:46:57 -0400
parents 21b5a9170b90
children
comparison
equal deleted inserted replaced
3:21b5a9170b90 4:f1eeaf42144b
27 reference_counters[reference]['antisense'][readlength] += 1 27 reference_counters[reference]['antisense'][readlength] += 1
28 else: 28 else:
29 reference_counters[reference]['sense'][readlength] += 1 29 reference_counters[reference]['sense'][readlength] += 1
30 df = pd.Panel(reference_counters).to_frame() 30 df = pd.Panel(reference_counters).to_frame()
31 df.index.names = ['readlength', 'orientation'] 31 df.index.names = ['readlength', 'orientation']
32 return df 32 return i, df
33 33
34 34
35 def to_long(df): 35 def to_long(df, total_count):
36 df = df.reset_index() 36 df = df.reset_index()
37 df = df.melt(id_vars=('readlength', 'orientation')) 37 df = df.melt(id_vars=('readlength', 'orientation'))
38 df.columns = ['readlength', 'orientation', 'reference', 'count'] 38 df.columns = ['readlength', 'orientation', 'reference', 'count']
39 df['TPM'] = df['count'] / total_count * 10 ** 6
39 return df 40 return df
40 41
41 42
42 def write_table(df, output_path): 43 def write_table(df, output_path):
43 df.to_csv(output_path, sep="\t", index=False) 44 df.to_csv(output_path, sep="\t", index=False)
50 @click.option('--wide/--long', default=False, help="Output wide or long format.") 51 @click.option('--wide/--long', default=False, help="Output wide or long format.")
51 @click.option('--output', default="/dev/stdout", help="Write to this file") 52 @click.option('--output', default="/dev/stdout", help="Write to this file")
52 def size_dist(alignment_path, minimum_size=18, maximum_size=30, output="/dev/stdout", wide=False): 53 def size_dist(alignment_path, minimum_size=18, maximum_size=30, output="/dev/stdout", wide=False):
53 """Calculate size distribution and orientation""" 54 """Calculate size distribution and orientation"""
54 with pysam.AlignmentFile(alignment_path) as alignment_file: 55 with pysam.AlignmentFile(alignment_path) as alignment_file:
55 df = global_size_distribution(alignment_file, minimum_size, maximum_size) 56 total_count, df = global_size_distribution(alignment_file, minimum_size, maximum_size)
56 if not wide: 57 if not wide:
57 df = to_long(df) 58 df = to_long(df, total_count)
58 write_table(df, output) 59 write_table(df, output)
59 60
60 61
61 if __name__ == '__main__': 62 if __name__ == '__main__':
62 size_dist() 63 size_dist()