From b9798355324f44a4a7d529aae4b1a7f842dd7bdc Mon Sep 17 00:00:00 2001
From: "Gervaise H. Henry" <gervaise.henry@utsouthwestern.edu>
Date: Mon, 25 Jan 2021 12:24:51 -0600
Subject: [PATCH] Convert strandedness from yes/no to t/f
---
CHANGELOG.md | 1 +
workflow/rna-seq.nf | 25 ++++++++++++++++++++++---
workflow/scripts/parse_meta.py | 7 +------
3 files changed, 24 insertions(+), 9 deletions(-)
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 637e6d9..9eb41cc 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,6 +1,7 @@
# v2.0.0
**User Facing**
* Endness metadata "Single Read" changed to "Single End" in data-hub, pipeline updated to handle (#110) ("Single Read" still acceptable for backwards compatibility)
+* Strandedness metadata "yes"/"no" changed to boolean "t"/"f" in data-hub, pipeline updated to handle (#70) ("yes"/"no" still acceptable for backwards compatibility)
**Background**
* Add memory limit (75%) per thread for samtools sort (#108)
diff --git a/workflow/rna-seq.nf b/workflow/rna-seq.nf
index bce0d20..57ec3ac 100644
--- a/workflow/rna-seq.nf
+++ b/workflow/rna-seq.nf
@@ -367,8 +367,27 @@ process parseMetadata {
echo -e "LOG: strandedness metadata parsed: \${stranded}" >> ${repRID}.parseMetadata.log
# get spike-in metadata
- spike=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p spike)
+ v=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p spike)
echo -e "LOG: spike-in metadata parsed: \${spike}" >> ${repRID}.parseMetadata.log
+ if [ "\${spike}" == "f" ]
+ then
+ spike="false"
+ elif [ "\${spike}" == "t" ]
+ then
+ spike="true"
+ elseif [ "\${spike}" == "no" ]
+ # "yes"/"no" depreciated as of Jan 2021, this option is present for backwards compatibility
+ then
+ spike="false"
+ elseif [ "\${spike}" == "yes" ]
+ # "yes"/"no" depreciated as of Jan 2021, this option is present for backwards compatibility
+ then
+ spike="true"
+ elif [ "\${spike}" == "nan" ]
+ then
+ endsRaw="_No value_"
+ endsMeta="NA"
+ fi
# get species metadata
species=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experiment}" -p species)
@@ -935,9 +954,9 @@ process inferMetadata {
# determine spike-in
if [ 1 -eq \$(echo \$(expr \${align_ercc} ">=" 10)) ]
then
- spike="yes"
+ spike="true"
else
- spike="no"
+ spike="false"
fi
echo -e "LOG: inference of strandedness results is: \${spike}" >> ${repRID}.inferMetadata.log
diff --git a/workflow/scripts/parse_meta.py b/workflow/scripts/parse_meta.py
index fdbc86c..b5379bb 100644
--- a/workflow/scripts/parse_meta.py
+++ b/workflow/scripts/parse_meta.py
@@ -73,12 +73,7 @@ def main():
# Get spike-in metadata from 'Experiment Settings.csv'
if (args.parameter == "spike"):
- if (metaFile.Used_Spike_Ins.unique() == "yes"):
- spike = "yes"
- elif (metaFile.Used_Spike_Ins.unique() == "no"):
- spike = "no"
- else:
- spike = metaFile.Used_Spike_Ins.unique()[0]
+ spike = metaFile.Used_Spike_Ins.unique()[0]
print(spike)
# Get species metadata from 'Experiment.csv'
--
GitLab