From 0158a2a031f0308480de4ac847054a56837be7ee Mon Sep 17 00:00:00 2001
From: Jeremy Mathews <Jeremy.Mathews@utsouthwestern.edu>
Date: Mon, 15 Jul 2019 09:21:16 -0500
Subject: [PATCH] test pool and psuedoreplicate for single experiment
---
workflow/scripts/pool_and_psuedoreplicate.py | 107 ++++++++++--------
.../tests/test_pool_and_psuedoreplicate.py | 8 ++
2 files changed, 65 insertions(+), 50 deletions(-)
diff --git a/workflow/scripts/pool_and_psuedoreplicate.py b/workflow/scripts/pool_and_psuedoreplicate.py
index 07eac44..93d1df4 100644
--- a/workflow/scripts/pool_and_psuedoreplicate.py
+++ b/workflow/scripts/pool_and_psuedoreplicate.py
@@ -172,58 +172,10 @@ def self_psuedoreplication(tag_file, prefix, paired):
return pseudoreplicate_dict
-def main():
- args = get_args()
- paired = args.paired
- design = args.design
- cutoff_ratio = args.cutoff
-
- # Create a file handler
- handler = logging.FileHandler('experiment_generation.log')
- logger.addHandler(handler)
-
- # Read files as dataframes
- design_df = pd.read_csv(design, sep='\t')
-
+def generate_design(design_df, replicated, single_control, pool_control, paired, cutoff_ratio):
# Get current directory to build paths
cwd = os.getcwd()
- # Check Number of replicates and replicates
- no_reps = check_replicates(design_df)
- no_unique_controls = check_controls(design_df)
-
- if no_reps == 1:
- logger.info("No other replicate specified "
- "so processing as an unreplicated experiment.")
- replicated = False
-
- else:
- logger.info("Multiple replicates specified "
- "so processing as a replicated experiment.")
- replicated = True
-
- if no_unique_controls == 1 and replicated:
- logger.info("Only a single control was specified "
- "so using same control for replicates, pool and psuedoreplicates.")
- single_control = True
- else:
- logger.info("Will merge only unique controls for pooled.")
- single_control = False
-
- # Pool the controls for checking
- if not single_control:
- control_df = get_read_count_ratio(design_df)
- control_files = design_df.control_tag_align.unique()
- pool_control = pool(control_files, "pool_control", paired)
- else:
- pool_control = design_df.control_tag_align.unique()[0]
-
- # if paired_end make tagAlign
- if paired:
- pool_control_tmp = bedpe_to_tagalign(pool_control, "pool_control")
- pool_control = pool_control_tmp
-
- # Psuedoreplicate and update design accordingly
if not replicated:
# Duplicate rows and update for pool and psuedoreplicates and update tagAlign with single end data
@@ -282,7 +234,6 @@ def main():
replicate_prefix = experiment_id + '_' + str(rep)
pr_dict = self_psuedoreplication(tag_file, replicate_prefix, paired)
pseudoreplicates_dict[rep] = pr_dict
-
# Merge self psuedoreplication for each true replicate
pseudoreplicates_df = pd.DataFrame.from_dict(pseudoreplicates_dict)
pool_pseudoreplicates_dict = {}
@@ -356,5 +307,61 @@ def main():
header=True, sep='\t', index=False)
+def main():
+ args = get_args()
+ paired = args.paired
+ design = args.design
+ cutoff_ratio = args.cutoff
+
+ # Create a file handler
+ handler = logging.FileHandler('experiment_generation.log')
+ logger.addHandler(handler)
+
+ # Read files as dataframes
+ design_df = pd.read_csv(design, sep='\t')
+
+ # Check Number of replicates and replicates
+ no_reps = check_replicates(design_df)
+ no_unique_controls = check_controls(design_df)
+
+ if no_reps == 1:
+ logger.info("No other replicate specified "
+ "so processing as an unreplicated experiment.")
+ replicated = False
+
+ else:
+ logger.info("Multiple replicates specified "
+ "so processing as a replicated experiment.")
+ replicated = True
+
+ if no_unique_controls == 1 and replicated:
+ logger.info("Only a single control was specified "
+ "so using same control for replicates, pool and psuedoreplicates.")
+ single_control = True
+ else:
+ logger.info("Will merge only unique controls for pooled.")
+ single_control = False
+
+ # Pool the controls for checking
+ if not single_control:
+ control_df = get_read_count_ratio(design_df)
+ control_files = design_df.control_tag_align.unique()
+l_df = get_read_count_ratio(design_df)
+ control_files = design_df.control_tag_align.unique()
+ pool_control = pool(control_files, "pool_control", paired)
+
+ pool_control = pool(control_files, "pool_control", paired)
+ else:
+ pool_control = design_df.control_tag_align.unique()[0]
+
+ # if paired_end make tagAlign
+ if paired:
+ pool_control_tmp = bedpe_to_tagalign(pool_control, "pool_control")
+ pool_control = pool_control_tmp
+
+ # Psuedoreplicate and update design accordingly
+ generate_design(design_df, replicated, single_control, pool_control, paired, cutoff_ratio)
+
+
if __name__ == '__main__':
main()
diff --git a/workflow/tests/test_pool_and_psuedoreplicate.py b/workflow/tests/test_pool_and_psuedoreplicate.py
index 31fffc5..58ceda9 100644
--- a/workflow/tests/test_pool_and_psuedoreplicate.py
+++ b/workflow/tests/test_pool_and_psuedoreplicate.py
@@ -60,6 +60,14 @@ def test_check_controls_single(design_experiment_3):
assert no_controls == 1
+@pytest.mark.unit
+def test_generate_design(design_experiment_2)
+ control_df = pool_and_psuedoreplicate.get_read_count_ratio(design_experiment_2)
+ control_files = design_experiment_2.control_tag_align.unique()
+ pool_control = pool_and_psuedoreplicate.pool(control_files, "pool_control", true)
+ new_design = pool_and_psuedoreplicate.generate_design(design_experiment_2, false, false, pool_control, true, 1.2)
+
+
@pytest.mark.singleend
def test_pool_and_psuedoreplicate_singleend():
design_file = os.path.join(test_output_path, 'ENCSR238SGC_ppr.tsv')
--
GitLab