From 59b0f72d0c712a7bbbb041b3d4dc695065072521 Mon Sep 17 00:00:00 2001
From: Laurent Modolo <laurent.modolo@ens-lyon.fr>
Date: Fri, 29 Mar 2019 13:27:06 +0100
Subject: [PATCH] hisat2: merge samtools and hisat2 in the process
---
src/nf_modules/hisat2/bam_converter.nf | 32 ------
src/nf_modules/hisat2/hisat2.config | 54 ----------
src/nf_modules/hisat2/hisat2.nf | 138 ------------------------
src/nf_modules/hisat2/indexing.config | 4 +-
src/nf_modules/hisat2/indexing.nf | 2 +-
src/nf_modules/hisat2/mapping_paired.nf | 17 ++-
src/nf_modules/hisat2/mapping_single.nf | 17 ++-
src/nf_modules/hisat2/tests.sh | 44 +++++---
8 files changed, 62 insertions(+), 246 deletions(-)
delete mode 100644 src/nf_modules/hisat2/bam_converter.nf
delete mode 100644 src/nf_modules/hisat2/hisat2.config
delete mode 100644 src/nf_modules/hisat2/hisat2.nf
diff --git a/src/nf_modules/hisat2/bam_converter.nf b/src/nf_modules/hisat2/bam_converter.nf
deleted file mode 100644
index f9627eb0..00000000
--- a/src/nf_modules/hisat2/bam_converter.nf
+++ /dev/null
@@ -1,32 +0,0 @@
-/*
-* SAMtools :
-* Imputs : sam files
-* Output : bam files
-*/
-
-/* sam to bam */
-params.sam = "$baseDir/data/bam/*.sam"
-
-log.info "sam files : ${params.sam}"
-
-Channel
- .fromPath( params.sam )
- .ifEmpty { error "Cannot find any sam files matching: ${params.sam}" }
- .set { sam_files }
-
-process bam_converter {
- tag "$sam"
- cpus 4
- publishDir "results/mapping/bam/", mode: 'copy'
-
- input:
- file sam from sam_files
-
- output:
- file "*.bam" into bam_files
-
- script:
-"""
-samtools view -@ ${task.cpus} -bS ${sam} > ${sam.baseName}.bam
-"""
-}
diff --git a/src/nf_modules/hisat2/hisat2.config b/src/nf_modules/hisat2/hisat2.config
deleted file mode 100644
index c2988690..00000000
--- a/src/nf_modules/hisat2/hisat2.config
+++ /dev/null
@@ -1,54 +0,0 @@
-profiles {
- docker {
- docker.temp = 'auto'
- docker.enabled = true
- process {
- $index_fasta {
- container = "hisat2:2.0.0"
- }
- $mapping_fastq {
- container = "hisat2:2.0.0"
- }
- $bam_converter {
- container = "samtools:1.7"
- }
- }
- }
- sge {
- process{
- $index_fasta {
- beforeScript = "module purge; module load Hisat2/2.0.0"
- executor = "sge"
- cpus = 1
- memory = "5GB"
- time = "6h"
- queueSize = 1000
- pollInterval = '60sec'
- queue = 'h6-E5-2667v4deb128'
- penv = 'openmp8'
- }
- $mapping_fastq {
- beforeScript = "module purge; module load Hisat2/2.0.0"
- executor = "sge"
- cpus = 4
- memory = "5GB"
- time = "6h"
- queueSize = 1000
- pollInterval = '60sec'
- queue = 'h6-E5-2667v4deb128'
- penv = 'openmp8'
- }
- $bam_converter {
- beforeScript = "module purge; module load SAMtools/1.5"
- executor = "sge"
- cpus = 4
- memory = "5GB"
- time = "6h"
- queueSize = 1000
- pollInterval = '60sec'
- queue = 'h6-E5-2667v4deb128'
- penv = 'openmp8'
- }
- }
- }
-}
diff --git a/src/nf_modules/hisat2/hisat2.nf b/src/nf_modules/hisat2/hisat2.nf
deleted file mode 100644
index e8d5eb63..00000000
--- a/src/nf_modules/hisat2/hisat2.nf
+++ /dev/null
@@ -1,138 +0,0 @@
-/*
-* Hisat2 :
-* Imputs : fastq files
-* Imputs : fasta files
-* Output : bam files
-*/
-
-/* fasta indexing */
-params.fasta = "$baseDir/data/bam/*.fasta"
-
-log.info "fasta files : ${params.fasta}"
-
-Channel
- .fromPath( params.fasta )
- .ifEmpty { error "Cannot find any fasta files matching: ${params.fasta}" }
- .set { fasta_file }
-
-process index_fasta {
- tag "$fasta.baseName"
- publishDir "results/mapping/index/", mode: 'copy'
-
- input:
- file fasta from fasta_file
-
- output:
- file "*.index*" into index_files
-
- script:
-"""
-hisat2-build ${fasta} ${fasta.baseName}.index
-"""
-}
-
-/*
-* for single-end data
-*/
-
-params.fastq = "$baseDir/data/fastq/*.fastq"
-params.index = "$baseDir/data/index/*.index*"
-
-log.info "fastq files : ${params.fastq}"
-log.info "index files : ${params.index}"
-
-Channel
- .fromPath( params.fastq )
- .ifEmpty { error "Cannot find any fastq files matching: ${params.fastq}" }
- .set { fastq_files }
-Channel
- .fromPath( params.index )
- .ifEmpty { error "Cannot find any index files matching: ${params.index}" }
- .set { index_files }
-
-process mapping_fastq {
- tag "$reads.baseName"
- cpus 4
- publishDir "results/mapping/", mode: 'copy'
-
- input:
- file reads from fastq_files
- file index from index_files.toList()
-
- output:
- file "*" into count_files
-
- script:
-"""
-hisat2 -x ${file(file(index[0]).baseName).baseName} -U ${reads} -S ${reads.baseName}.sam -p ${task.cpus}
-"""
-}
-
-/*
-* for paired-end data
-*/
-
-params.fastq = "$baseDir/data/fastq/*_{1,2}.fastq"
-params.index = "$baseDir/data/index/*.index.*"
-
-log.info "fastq files : ${params.fastq}"
-log.info "index files : ${params.index}"
-
-Channel
- .fromFilePairs( params.fastq )
- .ifEmpty { error "Cannot find any fastq files matching: ${params.fastq}" }
- .set { fastq_files }
-Channel
- .fromPath( params.index )
- .ifEmpty { error "Cannot find any index files matching: ${params.index}" }
- .set { index_files }
-
-process mapping_fastq {
- tag "$reads"
- //tag "$index.baseName"
- cpus 4
- publishDir "results/mapping/", mode: 'copy'
-
- input:
- set pair_id, file(reads) from fastq_files
- file index from index_files.toList()
-
- output:
- file "*" into counts_files
-
- script:
-"""
-hisat2 -x ${file(file(index[0]).baseName).baseName} -1 ${reads[0]} -2 ${reads[1]} -S ${pair_id}.sam -p ${task.cpus}
-"""
-}
-
-/*
-* converting sam into bam
-*/
-
-/* sam to bam */
-params.sam = "$baseDir/data/bam/*.sam"
-
-log.info "sam files : ${params.sam}"
-
-Channel
- .fromPath( params.sam )
- .ifEmpty { error "Cannot find any sam files matching: ${params.sam}" }
- .set { sam_files }
-
-process bam_converter {
- tag "$sam"
- cpus 4
- publishDir "results/mapping/bam/", mode: 'copy'
-
- input:
- file sam from sam_files
-
- output:
- file "*.bam" into bam_files
-
- script:
-"""
-samtools view -@ ${task.cpus} -bS ${sam} > ${sam.baseName}.bam
-"""
-}
diff --git a/src/nf_modules/hisat2/indexing.config b/src/nf_modules/hisat2/indexing.config
index dbbb4a2e..47c14a57 100644
--- a/src/nf_modules/hisat2/indexing.config
+++ b/src/nf_modules/hisat2/indexing.config
@@ -4,8 +4,8 @@ profiles {
docker.enabled = true
process {
withName: index_fasta {
- cpus = 4
container = "hisat2:2.0.0"
+ cpus = 4
}
}
}
@@ -18,7 +18,7 @@ profiles {
}
}
}
- sge {
+ psmn {
process{
withName: index_fasta {
beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
diff --git a/src/nf_modules/hisat2/indexing.nf b/src/nf_modules/hisat2/indexing.nf
index 563a13c2..1b11b3ef 100644
--- a/src/nf_modules/hisat2/indexing.nf
+++ b/src/nf_modules/hisat2/indexing.nf
@@ -27,6 +27,6 @@ process index_fasta {
script:
"""
-hisat2-build -p {task.cpus} ${fasta} ${fasta.baseName}.index
+hisat2-build -p ${task.cpus} ${fasta} ${fasta.baseName}.index
"""
}
diff --git a/src/nf_modules/hisat2/mapping_paired.nf b/src/nf_modules/hisat2/mapping_paired.nf
index d7bf9696..28b37e00 100644
--- a/src/nf_modules/hisat2/mapping_paired.nf
+++ b/src/nf_modules/hisat2/mapping_paired.nf
@@ -23,15 +23,26 @@ process mapping_fastq {
output:
file "*" into counts_files
+ set pair_id, "*.bam" into bam_files
+ file "*_report.txt" into mapping_report
script:
index_id = index[0]
for (index_file in index) {
- if (index_file =~ /.*\.1\.bt2/ && !(index_file =~ /.*\.rev\.1\.bt2/)) {
- index_id = ( index_file =~ /(.*)\.1\.bt2/)[0][1]
+ if (index_file =~ /.*\.1\.ht2/ && !(index_file =~ /.*\.rev\.1\.ht2/)) {
+ index_id = ( index_file =~ /(.*)\.1\.ht2/)[0][1]
}
}
"""
-hisat2 -x ${index_id} -1 ${reads[0]} -2 ${reads[1]} -S ${pair_id}.sam -p ${task.cpus}
+hisat2 -p ${task.cpus} \
+ -x ${index_id} \
+ -1 ${reads[0]} \
+ -2 ${reads[1]} 2> \
+${pair_id}_hisat2_report.txt | \
+samtools view -Sb - > ${pair_id}.bam
+
+if grep -q "Error" ${pair_id}_hisat2_report.txt; then
+ exit 1
+fi
"""
}
diff --git a/src/nf_modules/hisat2/mapping_single.nf b/src/nf_modules/hisat2/mapping_single.nf
index 93506140..0fdb729e 100644
--- a/src/nf_modules/hisat2/mapping_single.nf
+++ b/src/nf_modules/hisat2/mapping_single.nf
@@ -28,15 +28,26 @@ process mapping_fastq {
output:
file "*" into count_files
+ set file_id, "*.bam" into bam_files
+ file "*_report.txt" into mapping_report
script:
index_id = index[0]
for (index_file in index) {
- if (index_file =~ /.*\.1\.bt2/ && !(index_file =~ /.*\.rev\.1\.bt2/)) {
- index_id = ( index_file =~ /(.*)\.1\.bt2/)[0][1]
+ if (index_file =~ /.*\.1\.ht2/ && !(index_file =~ /.*\.rev\.1\.ht2/)) {
+ index_id = ( index_file =~ /(.*)\.1\.ht2/)[0][1]
}
}
"""
-hisat2 -x ${index_id} -U ${reads} -S ${file_id}.sam -p ${task.cpus}
+hisat2 -p ${task.cpus} \
+ -x ${index_id} \
+ -U ${reads} 2> \
+${file_id}_hisat2_report.txt | \
+samtools view -Sb - > ${file_id}.bam
+
+if grep -q "Error" ${file_id}_hisat2_report.txt; then
+ exit 1
+fi
+
"""
}
diff --git a/src/nf_modules/hisat2/tests.sh b/src/nf_modules/hisat2/tests.sh
index 261e9f2f..50e43966 100755
--- a/src/nf_modules/hisat2/tests.sh
+++ b/src/nf_modules/hisat2/tests.sh
@@ -1,21 +1,39 @@
-nextflow src/nf_modules/Hisat2/test/index.nf \
- -c src/nf_modules/Hisat2/hisat2.config \
+./nextflow src/nf_modules/hisat2/indexing.nf \
+ -c src/nf_modules/hisat2/indexing.config \
-profile docker \
- --fasta "data/tiny_dataset/fasta/tiny_v2.fasta"
+ --fasta "data/tiny_dataset/fasta/tiny_v2.fasta" \
+ -resume
-nextflow src/nf_modules/Hisat2/test/mapping_paired.nf \
- -c src/nf_modules/Hisat2/hisat2.config \
+./nextflow src/nf_modules/hisat2/mapping_paired.nf \
+ -c src/nf_modules/hisat2/mapping_paired.config \
-profile docker \
--index "results/mapping/index/tiny_v2.index*" \
- --fastq "data/tiny_dataset/fastq/tiny*_R{1,2}.fastq"
+ --fastq "data/tiny_dataset/fastq/tiny*_R{1,2}.fastq" \
+ -resume
-nextflow src/nf_modules/Hisat2/test/mapping_single.nf \
- -c src/nf_modules/Hisat2/hisat2.config \
+./nextflow src/nf_modules/hisat2/mapping_single.nf \
+ -c src/nf_modules/hisat2/mapping_single.config \
-profile docker \
--index "results/mapping/index/tiny_v2.index*" \
- --fastq "data/tiny_dataset/fastq/tiny*_S.fastq"
+ --fastq "data/tiny_dataset/fastq/tiny*_S.fastq" \
+ -resume
-nextflow src/nf_modules/Hisat2/test/bam_converter.nf \
- -c src/nf_modules/Hisat2/hisat2.config \
- -profile docker \
- --sam "results/mapping/*.sam" \
+if [ -x "$(command -v singularity)" ]; then
+./nextflow src/nf_modules/hisat2/indexing.nf \
+ -c src/nf_modules/hisat2/indexing.config \
+ -profile singularity \
+ --fasta "data/tiny_dataset/fasta/tiny_v2.fasta" \
+ -resume
+
+./nextflow src/nf_modules/hisat2/mapping_paired.nf \
+ -c src/nf_modules/hisat2/mapping_paired.config \
+ -profile singularity \
+ --index "results/mapping/index/tiny_v2.index*" \
+ --fastq "data/tiny_dataset/fastq/tiny*_R{1,2}.fastq"
+
+./nextflow src/nf_modules/hisat2/mapping_single.nf \
+ -c src/nf_modules/hisat2/mapping_single.config \
+ -profile singularity \
+ --index "results/mapping/index/tiny_v2.index*" \
+ --fastq "data/tiny_dataset/fastq/tiny*_S.fastq"
+fi
--
GitLab