From 4a26067455b3f4d38332d0d638189dd121d76d4e Mon Sep 17 00:00:00 2001
From: Dan Fornika <dfornika@gmail.com>
Date: Tue, 29 Aug 2023 12:28:23 -0700
Subject: [PATCH] Robust low coverage (#37)

* Handle low-coverage samples

* Fix duplicated sample ID
---
 bin/check_for_empty_assembly.py | 16 ++++++++++++++++
 modules/long_read_qc.nf         |  1 +
 modules/quast.nf                |  3 +++
 modules/unicycler.nf            |  1 +
 4 files changed, 21 insertions(+)
 create mode 100755 bin/check_for_empty_assembly.py

diff --git a/bin/check_for_empty_assembly.py b/bin/check_for_empty_assembly.py
new file mode 100755
index 0000000..09d3896
--- /dev/null
+++ b/bin/check_for_empty_assembly.py
@@ -0,0 +1,16 @@
+#!/usr/bin/env python3
+
+import argparse
+import os
+
+def main(args):
+    input_num_bytes = os.path.getsize(args.assembly)
+    if input_num_bytes == 0:
+        with open(args.assembly, 'w') as f:
+            f.write('>1\nN\n')
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--assembly', help='Input assembly to check', required=True)
+    args = parser.parse_args()
+    main(args)
diff --git a/modules/long_read_qc.nf b/modules/long_read_qc.nf
index a8ce0f5..4c3b745 100644
--- a/modules/long_read_qc.nf
+++ b/modules/long_read_qc.nf
@@ -80,6 +80,7 @@ process bandage {
     tag { sample_id + ' / ' + assembly_mode }
 
     executor 'local'
+    errorStrategy 'ignore'
 
     publishDir params.versioned_outdir ? "${params.outdir}/${sample_id}/${params.pipeline_short_name}-v${params.minor_version}-output" : "${params.outdir}/${sample_id}", pattern: "${sample_id}_${assembler}_${assembly_mode}_bandage.png", mode: 'copy'
 
diff --git a/modules/quast.nf b/modules/quast.nf
index 649d309..90cc93d 100644
--- a/modules/quast.nf
+++ b/modules/quast.nf
@@ -20,11 +20,14 @@ process quast {
       printf -- "          value: null\\n"                                                 >> ${sample_id}_${assembler}_${assembly_mode}_quast_provenance.yml
       printf -- "        - parameter: --fast\\n"                                           >> ${sample_id}_${assembler}_${assembly_mode}_quast_provenance.yml
       printf -- "          value: null\\n"                                                 >> ${sample_id}_${assembler}_${assembly_mode}_quast_provenance.yml
+      printf -- "        - parameter: --min-contig\\n"                                     >> ${sample_id}_${assembler}_${assembly_mode}_quast_provenance.yml
+      printf -- "          value: 0\\n"                                                    >> ${sample_id}_${assembler}_${assembly_mode}_quast_provenance.yml
 
       quast \
         --threads ${task.cpus} \
         --space-efficient \
         --fast \
+	--min-contig 0 \
         --output-dir ${sample_id} \
         ${assembly}
 
diff --git a/modules/unicycler.nf b/modules/unicycler.nf
index 0b24947..23c73cd 100644
--- a/modules/unicycler.nf
+++ b/modules/unicycler.nf
@@ -29,6 +29,7 @@ process unicycler {
 	${long_reads} \
         -o ${sample_id}_assembly
 
+      check_for_empty_assembly.py --assembly ${sample_id}_assembly/assembly.fasta
       sed 's/^>/>${sample_id}_/' ${sample_id}_assembly/assembly.fasta > ${sample_id}_unicycler_${assembly_mode}.fa
       cp ${sample_id}_assembly/assembly.gfa ${sample_id}_unicycler_${assembly_mode}.gfa
       cp ${sample_id}_assembly/unicycler.log ${sample_id}_unicycler_${assembly_mode}.log