Merge pull request #27 from CCBR/fix/issue24

Fix #23 and #24
CCBR · Feb 8, 2024 · 07b6554 · 07b6554
2 parents 82e2c0c + 91c6fff
commit 07b6554
Show file tree

Hide file tree

Showing 5 changed files with 54 additions and 6 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,12 @@
 ## Development Version
 
+## v1.0.1
+
+- feat: aggregate raw counts for hervquant (plus TPM) #24
+- fix: remove unnecessary normlization of locus and class aggregates (#23)
+- feat: version control doc mike plugin added
+- feat: GHA updates
+
 ## v1.0.0
 
 - feat: add annotation to HOMER output #12

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-1.0.0
+1.0.1
diff --git a/workflow/Snakefile b/workflow/Snakefile
@@ -4,8 +4,12 @@ include: join("rules","init.smk")
 def get_hervquant_output(wildcards):
     filelist = list()
     if GENOME.startswith("hg"):
-        filelist.extend(expand(join(WORKDIR,"results","hervquant","{replicate}","{replicate}.dummy"),replicate=REPLICATES))
+        filelist.extend(expand(join(WORKDIR,"results","hervquant","{replicate}","quant.sf"),replicate=REPLICATES))
         filelist.extend(expand(join(WORKDIR,"results","hervquant","{replicate}","{replicate}.Aligned.out.filtered.bam"),replicate=REPLICATES))
+        filelist.append(join(WORKDIR,"results","hervquant","hervquant_rawcounts.tsv"))
+        filelist.append(join(WORKDIR,"results","hervquant","hervquant_rawcounts.xlsx"))
+        filelist.append(join(WORKDIR,"results","hervquant","hervquant_TPM.tsv"))
+        filelist.append(join(WORKDIR,"results","hervquant","hervquant_TPM.xlsx"))
     return filelist
 
 localrules: all

diff --git a/workflow/rules/hervquant.smk b/workflow/rules/hervquant.smk
@@ -12,14 +12,14 @@ rule hervquant:
     input:
         unpack(get_hervquant_input),
     output:
-        dummy                       = join(WORKDIR,"results","hervquant","{replicate}","{replicate}.dummy"),
+        quantsf                     = join(WORKDIR,"results","hervquant","{replicate}","quant.sf"),
         filtered_bam                = join(WORKDIR,"results","hervquant","{replicate}","{replicate}.Aligned.out.filtered.bam"),
     params:
         name                        = "{replicate}",
         outdir                      = join(WORKDIR,"results","hervquant","{replicate}"),
         peorse                      = get_peorse,
         randomstr                   = str(uuid.uuid4()),
-        hervquantindexdir           = join(INDEXDIR,"hervquant"),
+        hervquantindexdir           = join(SCRIPTSDIR,"hervquant_aggregate.R"),
         hervquant_final_reference   = join(INDEXDIR,"hervquant","hervquant_final_reference.fa"),
     threads: getthreads("hervquant")
     envmodules: TOOLS["star"],
@@ -61,3 +61,40 @@ salmon quant \
  -o {params.outdir} \
  -p {threads}
 """
+
+rule hervquant_aggregate:
+    input:
+        expand(join(WORKDIR,"results","hervquant","{replicate}","quant.sf"),replicate=REPLICATES),
+    output:
+        rawcounts_tsv               = join(WORKDIR,"results","hervquant","hervquant_rawcounts.tsv"),
+        rawcounts_xlsx              = join(WORKDIR,"results","hervquant","hervquant_rawcounts.xlsx"),
+        TPM_tsv                     = join(WORKDIR,"results","hervquant","hervquant_TPM.tsv"),
+        TPM_xlsx                    = join(WORKDIR,"results","hervquant","hervquant_TPM.xlsx"),
+    params:
+        outdir                      = join(WORKDIR,"results","hervquant"),
+        randomstr                   = str(uuid.uuid4()),
+        hervquant_aggregate_rscript = join(SCRIPTSDIR,"hervquant_aggregate.R"),
+        hervquant_lookup            = join(INDEXDIR,"hervquant","herv_lookup.psv"),
+    threads: getthreads("hervquant")
+    envmodules: TOOLS["R"],
+    shell:"""
+{SETSTR}
+{TMPDIR_STR}
+
+cd {params.outdir}
+
+Rscript \\
+    {params.hervquant_aggregate_rscript} \\
+    --hervquantfolder {params.outdir} \\
+    --lookuptable {params.hervquant_lookup} \\
+    --outfile {output.rawcounts_tsv} \\
+    --exceloutfile {output.rawcounts_xlsx}
+
+Rscript \\
+    {params.hervquant_aggregate_rscript} \\
+    --hervquantfolder {params.outdir} \\
+    --lookuptable {params.hervquant_lookup} \\
+    --outfile {output.TPM_tsv} \\
+    --exceloutfile {output.TPM_xlsx} \\
+    --counttype TPM
+"""
diff --git a/workflow/rules/homer.smk b/workflow/rules/homer.smk
@@ -72,7 +72,7 @@ done
 analyzeRepeats.pl \\
     repeats \\
     {params.genome} \\
-    -d $list_of_dirs \\
+    -d $list_of_dirs -noadj \\
     > {output.counts}
 """
 
@@ -99,7 +99,7 @@ done
 analyzeRepeats.pl \\
     repeats \\
     {params.genome} \\
-    -d $list_of_dirs \\
+    -d $list_of_dirs -noadj \\
     > {output.counts}
 """