From 59506130ce458d56de83ef5e16c9823cb9630bfd Mon Sep 17 00:00:00 2001
From: Tobias Marschall <tobias.marschall@0ohm.net>
Date: Fri, 21 Sep 2018 16:39:07 +0200
Subject: [PATCH] Include statistics on complex regions in callset summary
statistics
---
Snakefile | 3 ++-
utils/callset_summary_stats.py | 42 ++++++++++++++++++++++++++++++++--
2 files changed, 42 insertions(+), 3 deletions(-)
diff --git a/Snakefile b/Snakefile
index e04e1f9..810e59a 100644
--- a/Snakefile
+++ b/Snakefile
@@ -942,6 +942,7 @@ rule summary_statistics:
segmentation = 'segmentation2/{sample}/{windows}.{bpdens}.txt',
strandstates = 'strand_states/{sample}/{windows}.{bpdens}/intitial_strand_state',
sv_calls = 'sv_calls/{sample}/{windows}.{bpdens}/{method}.txt',
+ complex = "sv_calls/{sample}/{windows}.{bpdens}/{method}.complex.tsv",
output:
tsv = 'stats/{sample}/{windows}.{bpdens,selected_j[0-9\\.]+_s[0-9\\.]+}/{method}.tsv',
log:
@@ -963,7 +964,7 @@ rule summary_statistics:
except KeyError:
pass
additional_params = ' '.join(p)
- shell('utils/callset_summary_stats.py --segmentation {input.segmentation} --strandstates {input.strandstates} {additional_params} {input.sv_calls} > {output.tsv} 2> {log}')
+ shell('utils/callset_summary_stats.py --segmentation {input.segmentation} --strandstates {input.strandstates} --complex-regions {input.complex} {additional_params} {input.sv_calls} > {output.tsv} 2> {log}')
rule aggregate_summary_statistics:
input:
diff --git a/utils/callset_summary_stats.py b/utils/callset_summary_stats.py
index 4711ad6..3142f78 100755
--- a/utils/callset_summary_stats.py
+++ b/utils/callset_summary_stats.py
@@ -44,6 +44,9 @@ def main():
help='Filename to read true clonal events from.')
parser.add_argument('--true-events-single-cell', default=None,
help='Filename to read true single-cell events from.')
+ parser.add_argument('--complex-regions', default=None,
+ help='Filename to read complex regions from.')
+
#parser.add_argument('--sce_min_distance', default=200000, type=int,
#help='Minimum distance of an SCE to a break in the joint segmentation.')
@@ -77,6 +80,22 @@ def main():
results['unique_calls'] = len(calls.groupby(by=['chrom','start','end','sv_call_name']))
result_order += ['total_calls','unique_calls']
+ # Read complex regions file (if provided)
+ if args.complex_regions is not None:
+ complex_regions = pd.read_csv(args.complex_regions, sep='\t')
+ assert set(complex_regions.columns) == set(['chrom','start','end'])
+ results['complex_lengths_mb'] = sum(complex_regions.end - complex_regions.start) / 1000000.0
+ result_order += ['complex_lengths_mb']
+ # determine which of the calls are complex
+ calls['is_complex'] = False
+ for chrom, start, end in complex_regions[['chrom','start','end']].values:
+ calls.loc[(calls.chrom==chrom) & (calls.start<end) & (calls.end>start), 'is_complex'] = True
+ results['total_calls_complex'] = len(calls[calls.is_complex])
+ results['unique_calls_complex'] = len(calls[calls.is_complex].groupby(by=['chrom','start','end','sv_call_name']))
+ result_order += ['total_calls_complex','unique_calls_complex']
+
+ print(calls, file=sys.stderr)
+
print('Detected {} total calls and {} unique calls in {} single cells'.format(
results['total_calls'],
results['unique_calls'],
@@ -86,9 +105,16 @@ def main():
# compute event lengths
calls = calls.assign(length=lambda x: x.end-x.start)
results['avg_sv_load_per_cell_mb'] = np.mean(calls.groupby(by=['cell']).length.sum()) / 1000000.0
+ result_order.append('avg_sv_load_per_cell_mb')
+ if 'is_complex' in calls.columns:
+ results['avg_sv_load_per_cell_complex_mb'] = np.mean(calls[calls.is_complex].groupby(by=['cell']).length.sum()) / 1000000.0
+ result_order.append('avg_sv_load_per_cell_complex_mb')
# create table of all unique SVs with the number of supporting cells
- sv_table = calls.groupby(by=['chrom','start','end','sv_call_name']).cell.count().reset_index(name='cell_count')
+ if 'is_complex' in calls.columns:
+ sv_table = calls.groupby(by=['chrom','start','end','sv_call_name','is_complex']).cell.count().reset_index(name='cell_count')
+ else:
+ sv_table = calls.groupby(by=['chrom','start','end','sv_call_name']).cell.count().reset_index(name='cell_count')
sv_table = sv_table.assign(af=lambda x: x.cell_count/results['cell_count'])
sv_table = sv_table.assign(length=lambda x: x.end-x.start)
@@ -114,7 +140,19 @@ def main():
results['length_sum_af10to80_mb'] = sv_table[(sv_table.af>=0.1) & (sv_table.af<0.8)].length.sum() / 1000000.0
results['length_sum_af80to100_mb'] = sv_table[sv_table.af>=0.8].length.sum() / 1000000.0
- result_order += ['avg_sv_load_per_cell_mb','calls_af0to10','calls_af10to80','calls_af80to100','length_sum_af0to10_mb','length_sum_af10to80_mb','length_sum_af80to100_mb']
+ result_order += ['calls_af0to10','calls_af10to80','calls_af80to100','length_sum_af0to10_mb','length_sum_af10to80_mb','length_sum_af80to100_mb']
+
+ if 'is_complex' in calls.columns:
+ results['calls_af0to10_complex'] = len(sv_table[sv_table.is_complex & (sv_table.af<0.1)])
+ results['calls_af10to80_complex'] = len(sv_table[sv_table.is_complex & (sv_table.af>=0.1) & (sv_table.af<0.8)])
+ results['calls_af80to100_complex'] = len(sv_table[sv_table.is_complex & (sv_table.af>=0.8)])
+
+ results['length_sum_af0to10_complex_mb'] = sv_table[sv_table.is_complex & (sv_table.af<0.1)].length.sum() / 1000000.0
+ results['length_sum_af10to80_complex_mb'] = sv_table[sv_table.is_complex & (sv_table.af>=0.1) & (sv_table.af<0.8)].length.sum() / 1000000.0
+ results['length_sum_af80to100_complex_mb'] = sv_table[sv_table.is_complex & (sv_table.af>=0.8)].length.sum() / 1000000.0
+ result_order += ['calls_af0to10_complex','calls_af10to80_complex','calls_af80to100_complex','length_sum_af0to10_complex_mb','length_sum_af10to80_complex_mb','length_sum_af80to100_complex_mb']
+
+
print(*result_order, sep='\t')
print(*(results[x] for x in result_order), sep='\t')
--
GitLab