From 95f70fe7a1ba9eacb086560f075d8db2a561a0ee Mon Sep 17 00:00:00 2001
From: Valentina Galata <valentina.galata@uni.lu>
Date: Thu, 2 Jul 2020 12:58:39 +0200
Subject: [PATCH] workflow: rm partial gene counts rule from analysis (issue
#50)
---
workflow/rules/analysis.smk | 26 --------------------------
workflow/steps/analysis.smk | 4 ----
2 files changed, 30 deletions(-)
diff --git a/workflow/rules/analysis.smk b/workflow/rules/analysis.smk
index 07b63ec..b457cd9 100644
--- a/workflow/rules/analysis.smk
+++ b/workflow/rules/analysis.smk
@@ -26,23 +26,6 @@ rule analysis_quast:
##################################################
# Proteins
-# Number of partial genes
-rule analysis_prodigal_partial:
- input:
- os.path.join(RESULTS_DIR, "annotation/prodigal/{rtype}/{tool}/proteins.faa")
- output:
- os.path.join(RESULTS_DIR, "annotation/prodigal/{rtype}/{tool}/partial_gene_counts.txt")
- wildcard_constraints:
- rtype="|".join(READ_TYPES),
- tool="|".join(ASSEMBLERS)
- threads: 1
- message:
- "Count partial genes in {input}"
- shell:
- "grep -E -o 'partial.{{0,3}}' {input} | grep '10\|01\|11' | wc -l > {output}"
-
-# TODO: summary for tables/figures: separate *.smk
-
# for CD-HIT
rule analysis_bbmap_rename:
input:
@@ -93,15 +76,6 @@ rule analysis_cdhit:
"cd-hit-2d -i {input.faa2} -i2 {input.faa1} -o {output.faa21} -c 0.9 -n 5 -d 0 -M 16000 -T 8 && "
"date) 2> {log.err} > {log.out}"
-# TODO: summary for tables/figures: separate *.smk
-# """
-# cd-hit-2d -i {input.co1} -i2 {input.co2} -o {output.form1} -c 0.9 -n 5 -d 0 -M 16000 -T 8 &&\
-# cd-hit-2d -i {input.co2} -i2 {input.co1} -o {output.form2} -c 0.9 -n 5 -d 0 -M 16000 -T 8 &&\
-# echo {output.form1} > tmp && grep -c '>' {output.form1} >> tmp &&\
-# echo {output.form2} >> tmp && grep -c '>' {output.form2} >> tmp &&\
-# cat tmp | paste - - > {output.form3}
-# """
-
##################################################
# MMseqs2
diff --git a/workflow/steps/analysis.smk b/workflow/steps/analysis.smk
index 144d642..b72948c 100644
--- a/workflow/steps/analysis.smk
+++ b/workflow/steps/analysis.smk
@@ -24,10 +24,6 @@ rule ANALYSIS_ASSEMBLY:
rule ANALYSIS_PROTEINS:
input:
- expand(
- os.path.join(RESULTS_DIR, "annotation/prodigal/{rtype_tool}/partial_gene_counts.txt"),
- rtype_tool=["%s/%s" % (rtype, tool) for rtype, tool in READ_ASSEMBLERS]
- ),
expand(
os.path.join(RESULTS_DIR, "analysis/cdhit/{combi}.faa"),
combi=["%s_%s__%s_%s" % (p[0][0], p[0][1], p[1][0], p[1][1]) for p in READ_ASSEMBLER_PAIRS]
--
GitLab