Merge pull request #155 from sjackman/bandage-gfa

abyss, bandage, spades: Add GFA

Author: ODiogoSilva

changelog.md

@@ -13,6 +13,9 @@ recipes, the `--recipe-list` and `--recipe-list-short` options were added.
 
 - Added new `disableRR` param in the `spades` component that disables repeat
 resolution
+- The `abyss` and `spades` components emit GFA in a secondary channel.
+- The new `bandage` component can accept either FASTA from a primary channel
+  or GFA from a secondary channel.
 
 ### New components
 

flowcraft/generator/components/assembly.py

@@ -27,6 +27,7 @@ def __init__(self, **kwargs):
         self.output_type = "fasta"
 
         self.link_end.append({"link": "SIDE_max_len", "alias": "SIDE_max_len"})
+        self.link_start.append("gfa1")
 
         self.dependencies = ["integrity_coverage"]
 
@@ -200,6 +201,7 @@ def __init__(self, **kwargs):
 
         self.input_type = "fastq"
         self.output_type = "fasta"
+        self.link_start.append("gfa1")
 
         self.params = {
             "abyssKmer": {

flowcraft/generator/components/assembly_processing.py

@@ -247,6 +247,8 @@ def __init__(self, **kwargs):
         self.input_type = "fasta"
         self.output_type = None
 
+        self.link_end.append({"link": "gfa1", "alias": "gfa1"})
+
         self.params = {
             "reference": {
                 "default": "null",

flowcraft/generator/templates/abyss.nf

@@ -1,26 +1,23 @@
-
-IN_abyss_kmer_{{ pid }} = Channel.value(params.abyssKmer{{ param_id }})
-
 process abyss_{{ pid }} {
-    
-    // Send POST request to platform
     {% include "post.txt" ignore missing %}
 
     tag { sample_id }
-    publishDir 'results/assembly/abyss_{{ pid }}/', pattern: '*-scaffolds.fa', mode: 'copy'
+    publishDir 'results/assembly/abyss_{{ pid }}/', pattern: '*-scaffolds.fa'
+    publishDir 'results/assembly/abyss_{{ pid }}/', pattern: '*-scaffolds.gfa'
 
     input:
     set sample_id, file(fastq_pair) from {{ input_channel }}
-    val kmer from IN_abyss_kmer_{{ pid }}
+    val k from Channel.value(params.abyssKmer{{ param_id }})
 
     output:
-    set sample_id, file('*.fa') into {{ output_channel }}
+    set sample_id, file('*-scaffolds.fa') into {{ output_channel }}
+    file "*-scaffolds.gfa" into gfa1_{{ pid }}
     {% with task_name="abyss" %}
     {%- include "compiler_channels.txt" ignore missing -%}
     {% endwith %}
 
     script:
-    "abyss-pe in=\"${fastq_pair[0]} ${fastq_pair[1]}\" k=${kmer} name=${sample_id}"
+    "abyss-pe name=${sample_id} graph=gfa k=${k} v=-v in=\"${fastq_pair[0]} ${fastq_pair[1]}\""
 }
 
 {{ forks }}

flowcraft/generator/templates/bandage.nf

@@ -1,50 +1,39 @@
-if (params.reference{{param_id}} != null) {
-    process bandage_{{pid}} {
-        {% include "post.txt" ignore missing %}
-
-        tag { sample_id }
-        publishDir "reports/assembly/bandage_{{pid}}/$sample_id"
-
-        input:
-        set sample_id, file(assembly) from {{input_channel}}
-        file reference from Channel.fromPath(params.reference{{param_id}})
-
-        output:
-        file "*.png"
-        file "*.svg"
-        {% with task_name="bandage" %}
-        {%- include "compiler_channels.txt" ignore missing -%}
-        {% endwith %}
-
-        script:
-        """
-        time Bandage image $assembly ${assembly}.png >>.command.log 2>&1
-        time Bandage image $assembly ${assembly}.svg >>.command.log 2>&1
-        time Bandage image $assembly ${assembly}.ref.png --query $reference >>.command.log 2>&1
-        time Bandage image $assembly ${assembly}.ref.svg --query $reference >>.command.log 2>&1
-        """
-    }
-} else {
-    process bandage_{{pid}} {
-        {% include "post.txt" ignore missing %}
-
-        tag { sample_id }
-        publishDir "reports/assembly/bandage_{{pid}}/$sample_id"
-
-        input:
-        set sample_id, file(assembly) from {{input_channel}}
-
-        output:
-        file "*.png"
-        file "*.svg"
-        {% with task_name="bandage" %}
-        {%- include "compiler_channels.txt" ignore missing -%}
-        {% endwith %}
-
-        script:
+// True when a GFA secondary channel is connected to this component.
+has_gfa1_{{pid}} = binding.hasVariable('gfa1_{{pid}}')
+
+process bandage_{{pid}} {
+    {% include "post.txt" ignore missing %}
+
+    tag { sample_id }
+    publishDir "reports/assembly/bandage_{{pid}}/$sample_id"
+
+    input:
+    set sample_id, file(fasta) from {{input_channel}}
+    file gfa1 from has_gfa1_{{pid}} ? gfa1_{{pid}} : Channel.value("NA")
+    file reference from params.reference{{param_id}} ?
+        Channel.fromPath(params.reference{{param_id}}) :
+        Channel.value("NA")
+
+    output:
+    file "*.png"
+    file "*.svg"
+    {% with task_name="bandage" %}
+    {%- include "compiler_channels.txt" ignore missing -%}
+    {% endwith %}
+
+    script:
+    // Use the GFA assembly when available and FASTA otherwise.
+    assembly = has_gfa1_{{pid}} ? gfa1 : fasta
+    command =
         """
         time Bandage image $assembly ${assembly}.png >>.command.log 2>&1
         time Bandage image $assembly ${assembly}.svg >>.command.log 2>&1
         """
-    }
+    if (params.reference{{param_id}})
+        command +=
+            """
+            time Bandage image $assembly ${assembly}.ref.png --query $reference >>.command.log 2>&1
+            time Bandage image $assembly ${assembly}.ref.svg --query $reference >>.command.log 2>&1
+            """
+    command
 }

flowcraft/generator/templates/spades.nf

@@ -43,7 +43,7 @@ process spades_{{ pid }} {
     output:
     set sample_id, file('*_spades*.fasta') into {{ output_channel }}
     file "*.fastg" optional true
-    file "*.gfa" optional true
+    file "*.gfa" into gfa1_{{ pid }}
     {% with task_name="spades" %}
     {%- include "compiler_channels.txt" ignore missing -%}
     {% endwith %}