let cnspath inherit labels from existing assembly, propagate coverage…

…s/graphs in cnspath, fixes issues #241 and #260
marbl · Jun 13, 2024 · 4aeed14 · 4aeed14
1 parent 5e1cbd7
commit 4aeed14
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 2 deletions.
diff --git a/src/Snakefile b/src/Snakefile
@@ -154,13 +154,19 @@ fi
 
 rule cnspath:
     input:
+        graph       = '5-untip/unitig-unrolled-unitig-unrolled-popped-unitig-normal-connected-tip.gfa',
+        hifi5cov    = '5-untip/unitig-unrolled-unitig-unrolled-popped-unitig-normal-connected-tip.hifi-coverage.csv',
+        ont5cov     = '5-untip/unitig-unrolled-unitig-unrolled-popped-unitig-normal-connected-tip.ont-coverage.csv',
         layout      = '6-layoutContigs/unitig-popped.layout',
         scfmap      = '6-layoutContigs/unitig-popped.layout.scfmap',
         consensus   = '7-consensus/unitig-popped.fasta',
         cnshap1     = '7-consensus/unitig-popped.haplotype1.fasta',
         cnshap2     = '7-consensus/unitig-popped.haplotype2.fasta',
         cnsunas     = '7-consensus/unitig-popped.unassigned.fasta'
     output:
+        graph       = 'assembly.homopolymer-compressed.gfa',
+        hificov     = 'assembly.hifi-coverage.csv',
+        ontcov      = 'assembly.ont-coverage.csv',
         layout      = 'assembly.homopolymer-compressed.layout',
         consensus   = 'assembly.fasta',
         scfmap      = 'assembly.scfmap'
@@ -170,10 +176,15 @@ rule cnspath:
         cnsunas     = 'assembly.unassigned.fasta',
     shell:
         '''
+cp -p {input.graph}     {output.graph}
 cp -p {input.layout}    {output.layout}
+cp -p {input.hifi5cov}  {output.hificov}
+cp -p {input.ont5cov}   {output.ontcov}
 cp -p {input.consensus} {output.consensus}
-cp -p {input.cnshap1}   {params.cnshap1}
-cp -p {input.cnshap2}   {params.cnshap2}
+if [  -s {input.cnshap1} -a -s {input.cnshap2} ]; then
+  cp -p {input.cnshap1}   {params.cnshap1}
+  cp -p {input.cnshap2}   {params.cnshap2}
+fi
 cp -p {input.cnsunas}   {params.cnsunas}
 cp -p {input.scfmap}    {output.scfmap}
         '''

diff --git a/src/verkko.sh b/src/verkko.sh
@@ -1207,6 +1207,7 @@ if [ "x$cnspaths" != "x" ] ; then
     #  Copy pieces from the previous assembly to the new run directory.  This
     #  is done - instead of symlinking - to prevent Snakemake from
     #  'accidentally' obliterating precious original files.
+    cp -p ${cnsassembly}/assembly.homopolymer-compressed.noseq.gfa                               ${outd}/assembly.homopolymer-compressed.noseq.gfa
 
     if [ ! -e "${outd}/5-untip" ]; then
        mkdir ${outd}/5-untip
@@ -1232,6 +1233,19 @@ if [ "x$cnspaths" != "x" ] ; then
         cp -p ${cnsassembly}/7-consensus/ont_subset.fasta.gz          ${outd}/7-consensus/ont_subset.fasta.gz
         cp -p ${cnsassembly}/7-consensus/ont_subset.id                ${outd}/7-consensus/ont_subset.id
     fi
+
+    # copy haplotype labels if they exist
+    if [ -e ${cnsassembly}/6-rukki/label1 ]; then
+       mkdir ${outd}/6-rukki
+       cp -p ${cnsassembly}/6-rukki/label1                           ${outd}/6-rukki/label1
+       cp -p ${cnsassembly}/6-rukki/label2                           ${outd}/6-rukki/label2
+    fi
+    if [ -e ${cnsassembly}/8-hicPipeline/label1 ]; then
+       mkdir ${outd}/6-rukki
+       cp -p ${cnsassembly}/8-hicPipeline/label1                     ${outd}/6-rukki/label1
+       cp -p ${cnsassembly}/8-hicPipeline/label2                     ${outd}/6-rukki/label2
+    fi
+
     cp -p ${cnsassembly}/emptyfile ${outd}/emptyfile
 fi
 
@@ -1332,6 +1346,7 @@ if [ "x$withhic" = "xTrue" -o "x$withporec" = "xTrue" ] ; then
     ret=$?
     if [ $ret -eq 0 ]; then
         cp *.fasta ../../
+        cp *.gfa ../../
         cp *.layout ../../
         cp *.scfmap ../../
         cp ../rukki.paths.tsv ../../assembly.paths.tsv