diamond blast should run with the --taxon-exclude option

sanger-tol · Jun 24, 2024 · cfac699 · cfac699
1 parent 2ae99f6
commit cfac699
Show file tree

Hide file tree

Showing 8 changed files with 74 additions and 5 deletions.
diff --git a/modules.json b/modules.json
@@ -30,12 +30,14 @@
                     "diamond/blastp": {
                         "branch": "master",
                         "git_sha": "b29f6beb86d1d24d680277fb1a3f4de7b8b8a92c",
-                        "installed_by": ["modules"]
+                        "installed_by": ["modules"],
+                        "patch": "modules/nf-core/diamond/blastp/diamond-blastp.diff"
                     },
                     "diamond/blastx": {
                         "branch": "master",
                         "git_sha": "b29f6beb86d1d24d680277fb1a3f4de7b8b8a92c",
-                        "installed_by": ["modules"]
+                        "installed_by": ["modules"],
+                        "patch": "modules/nf-core/diamond/blastx/diamond-blastx.diff"
                     },
                     "fastawindows": {
                         "branch": "master",

diff --git a/modules/nf-core/diamond/blastp/diamond-blastp.diff b/modules/nf-core/diamond/blastp/diamond-blastp.diff
diff --git a/modules/nf-core/diamond/blastp/main.nf b/modules/nf-core/diamond/blastp/main.nf
diff --git a/modules/nf-core/diamond/blastx/diamond-blastx.diff b/modules/nf-core/diamond/blastx/diamond-blastx.diff
diff --git a/modules/nf-core/diamond/blastx/main.nf b/modules/nf-core/diamond/blastx/main.nf
diff --git a/subworkflows/local/busco_diamond_blastp.nf b/subworkflows/local/busco_diamond_blastp.nf
@@ -39,6 +39,7 @@ workflow BUSCO_DIAMOND {
     | transpose()
     | filter { rank,id -> rank =~ /species/ }
     | map { rank, id -> id}
+    | first
     | set { ch_taxid }
 
 
@@ -116,7 +117,7 @@ workflow BUSCO_DIAMOND {
     // Hardcoded to match the format expected by blobtools
     def outext = 'txt'
     def cols   = 'qseqid staxids bitscore qseqid sseqid pident length mismatch gapopen qstart qend sstart send evalue bitscore'
-    DIAMOND_BLASTP ( ch_busco_genes, blastp, outext, cols )
+    DIAMOND_BLASTP ( ch_busco_genes, blastp, outext, cols, ch_taxid )
     ch_versions = ch_versions.mix ( DIAMOND_BLASTP.out.versions.first() )
 
 

diff --git a/subworkflows/local/run_blastx.nf b/subworkflows/local/run_blastx.nf
@@ -11,6 +11,7 @@ workflow RUN_BLASTX {
     fasta      // channel: [ val(meta), path(fasta) ]
     table      // channel: [ val(meta), path(busco_table) ]
     blastx     // channel: [ val(meta), path(blastx_db) ]
+    taxon_id   // channel: val(taxon_id)
 
 
     main:
@@ -30,7 +31,7 @@ workflow RUN_BLASTX {
     // Hardocded to match the format expected by blobtools
     def outext = 'txt'
     def cols   = 'qseqid staxids bitscore qseqid sseqid pident length mismatch gapopen qstart qend sstart send evalue bitscore'
-    DIAMOND_BLASTX ( BLOBTOOLKIT_CHUNK.out.chunks, blastx, outext, cols)
+    DIAMOND_BLASTX ( BLOBTOOLKIT_CHUNK.out.chunks, blastx, outext, cols, taxon_id )
     ch_versions = ch_versions.mix ( DIAMOND_BLASTX.out.versions.first() )
 
 

diff --git a/workflows/blobtoolkit.nf b/workflows/blobtoolkit.nf
@@ -140,6 +140,7 @@ workflow BLOBTOOLKIT {
         PREPARE_GENOME.out.genome,
         BUSCO_DIAMOND.out.first_table,
         ch_blastx,
+        BUSCO_DIAMOND.out.taxon_id,
     )
     ch_versions = ch_versions.mix ( RUN_BLASTX.out.versions )
 
@@ -151,7 +152,7 @@ workflow BLOBTOOLKIT {
         RUN_BLASTX.out.blastx_out,
         PREPARE_GENOME.out.genome,
         ch_blastn,
-        BUSCO_DIAMOND.out.taxon_id
+        BUSCO_DIAMOND.out.taxon_id,
     )
 
     //