Apply suggestions from code review + add support for GFF annotations

modules/antismash/antismashlite/main.nf

@@ -17,6 +17,7 @@ process ANTISMASH_ANTISMASHLITE {
 
     input:
     tuple val(meta), path(sequence_input)
+    path(gff)
     path(databases)
     path(antismash_dir) // Optional input: AntiSMASH installation folder. It is not needed for using this module with conda, but required for docker/singularity (see meta.yml).
 
@@ -46,6 +47,12 @@ process ANTISMASH_ANTISMASHLITE {
     script:
     def args = task.ext.args ?: ''
     prefix = task.ext.suffix ? "${meta.id}${task.ext.suffix}" : "${meta.id}"
+    if ( sequence_input.getExtension != 'fasta' && sequence_input.getExtension != 'fna' && gff )
+        log.warn "GFF input to antiSMASH can only be used if FASTA sequence input is supplied. GFF will be ignored for sample ${meta.id}"
+    if ( (sequence_input.getExtension == 'fasta' || sequence_input.getExtension == 'fna') && gff )
+        gff_flag = "--genefinding-gff3 ${gff}"
+    else
+        gff_flag = ""
 
     """
     ## We specifically do not include annotations (--genefinding-tool none) as
@@ -57,6 +64,7 @@ process ANTISMASH_ANTISMASHLITE {
         --genefinding-tool none \\
         --logfile $prefix/${prefix}.log \\
         --databases $databases \\
+        $gff_flag \\
         $sequence_input
 
     cat <<-END_VERSIONS > versions.yml

modules/antismash/antismashlite/meta.yml

@@ -34,21 +34,24 @@ input:
   - sequence_input:
       type: file
       description: nucleotide sequence file (annotated)
-      pattern: "*.{gbk, gb, gbff, genbank, embl}"
+      pattern: "*.{gbk, gb, gbff, genbank, embl, fasta, fna}"
+  - gff:
+      type: file
+      description: GFF3 file to extract annotated features from (only needed if sequence_input is in FASTA format)
+      pattern: "*.{gff, GFF}"
   - databases:
       type: directory
-      description: downloaded antismash databases e.g. data/databases
-      pattern: "*"
+      description: downloaded AntiSMASH databases e.g. data/databases
+      pattern: "*/"
   - antismash_dir:
       type: directory
       description: |
-        antismash installation folder which is being modified during the
-        antiSMASH database downloading step. The modified files are
-        normally downloaded by download-antismash-databases itself, and must be
-        retrieved by the user by manually running the command within antismash-lite.
-        As this folder cannot be modified within the docker and singularity containers,
-        it needs to be mounted (including all modified files from the downloading step)
-        as a workaround.
+        A local copy of an AntiSMASH installation folder. This is required when running with
+        docker and singularity (not required for conda), due to attempted 'modifications' of 
+        files during database checks in the installation directory, something that cannot 
+        be done in immutable docker/singularity containers. Therefore, a local installation
+        directory needs to be mounted (including all modified files from the downloading step)
+        to the container as a workaround.
       pattern: "*"
 
 output:
@@ -87,11 +90,11 @@ output:
       pattern: "knownclusterblast/*_c*.txt"
   - svg_files_clusterblast:
       type: file
-      description: SVG image showing the % identity of the aligned hits against their queries
+      description: SVG images showing the % identity of the aligned hits against their queries
       pattern: "svg/clusterblast*.svg"
   - svg_files_knownclusterblast:
       type: file
-      description: SVG image showing the % identity of the aligned hits against their queries
+      description: SVG images showing the % identity of the aligned hits against their queries
       pattern: "svg/knownclusterblast*.svg"
   - gbk_input:
       type: file
@@ -127,7 +130,7 @@ output:
       pattern: "knownclusterblastoutput.txt"
   - json_sideloading:
       type: file
-      description: Sideloaded annotations of protoclusters and/or subregions (see documentation "Annotation sideloading")
+      description: Sideloaded annotations of protoclusters and/or subregions (see antiSMASH documentation "Annotation sideloading")
       pattern: "regions.js"
 
 authors:

tests/modules/antismash/antismashlite/main.nf

@@ -32,5 +32,5 @@ workflow test_antismashlite {
     UNTAR2 ( input_antismash_db2 )
     UNTAR3 ( input_antismash_db3 )
     ANTISMASH_ANTISMASHLITEDOWNLOADDATABASES ( UNTAR1.out.untar.map{ it[1] },  UNTAR2.out.untar.map{ it[1] }, UNTAR3.out.untar.map{ it[1] } )
-    ANTISMASH_ANTISMASHLITE ( input_genome, ANTISMASH_ANTISMASHLITEDOWNLOADDATABASES.out.database, ANTISMASH_ANTISMASHLITEDOWNLOADDATABASES.out.antismash_dir )
+    ANTISMASH_ANTISMASHLITE ( input_genome, ANTISMASH_ANTISMASHLITEDOWNLOADDATABASES.out.database, ANTISMASH_ANTISMASHLITEDOWNLOADDATABASES.out.antismash_dir, [] )
 }