Merge branch 'main' into samtools_sort

CHANGELOG.md

@@ -36,6 +36,8 @@
     - `salmon/salmon_quant`: Transcript quantification from RNA-seq data (PR #24).
 
 * `samtools`:
+    - `samtools/flagstat`: Counts the number of alignments in SAM/BAM/CRAM files for each FLAG type (PR #31).
+    - `samtools/idxstats`: Reports alignment summary statistics for a SAM/BAM/CRAM file (PR #32).    
     - `samtools/samtools_sort`: Sort SAM/BAM/CRAM files (PR #36).
 
 ## MAJOR CHANGES

src/gffread/test.sh

@@ -5,9 +5,9 @@
 
 set -e
 
-test_output_dir="${meta_resources_dir}test_data/test_output"
-test_dir="${meta_resources_dir}test_data"
-expected_output_dir="${meta_resources_dir}test_data/output"
+test_output_dir="${meta_resources_dir}/test_data/test_output"
+test_dir="${meta_resources_dir}/test_data"
+expected_output_dir="${meta_resources_dir}/test_data/output"
 
 mkdir -p "$test_output_dir"
 
@@ -108,4 +108,4 @@ rm -r "$test_output_dir"
 
 echo "> All tests successful"
 
-exit 0
+exit 0

src/samtools/samtools_flagstat/config.vsh.yaml

@@ -0,0 +1,51 @@
+name: samtools_flagstat
+namespace: samtools
+description: Counts the number of alignments in SAM/BAM/CRAM files for each FLAG type.
+keywords: [ stats, mapping, counts, bam, sam, cram ]
+links:
+  homepage: https://www.htslib.org/
+  documentation: https://www.htslib.org/doc/samtools-flagstat.html
+  repository: https://github.com/samtools/samtools
+references: 
+  doi: [10.1093/bioinformatics/btp352, 10.1093/gigascience/giab008]
+license: MIT/Expat
+
+argument_groups:
+  - name: Inputs
+    arguments:
+    - name: --bam
+      type: file
+      description: |
+        BAM input files.
+    - name: --bai
+      type: file
+      description: |
+        BAM index file.
+  - name: Outputs 
+    arguments:
+    - name: --output
+      type: file 
+      description: |
+        File containing samtools stats output.
+      direction: output
+      example: output.flagstat
+
+resources:
+  - type: bash_script
+    path: script.sh
+test_resources:
+  - type: bash_script
+    path: test.sh
+  - type: file
+    path: test_data
+engines:
+  - type: docker
+    image: quay.io/biocontainers/samtools:1.19.2--h50ea8bc_1
+    setup:
+      - type: docker
+        run: |
+          samtools --version 2>&1 | grep -E '^(samtools|Using htslib)' | \
+          sed 's#Using ##;s# \([0-9\.]*\)$#: \1#' > /var/software_versions.txt
+runners:
+- type: executable
+- type: nextflow

src/samtools/samtools_flagstat/help.txt

@@ -0,0 +1,13 @@
+```sh
+samtools flagstat --help
+```
+Usage: samtools flagstat [options] <in.bam>
+      --input-fmt-option OPT[=VAL]
+               Specify a single input file format option in the form
+               of OPTION or OPTION=VALUE
+  -@, --threads INT
+               Number of additional threads to use [0]
+      --verbosity INT
+               Set level of verbosity
+  -O, --output-fmt FORMAT[,OPT[=VAL]]...
+               Specify output format (json, tsv)

src/samtools/samtools_flagstat/script.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+
+## VIASH START
+## VIASH END
+
+set -e
+
+samtools flagstat \
+    "$par_bam" \
+    > "$par_output"
+

src/samtools/samtools_flagstat/test.sh

@@ -0,0 +1,47 @@
+#!/bin/bash
+
+test_dir="${meta_resources_dir}/test_data"
+echo ">>> Testing $meta_functionality_name"
+
+"$meta_executable" \
+  --bam "$test_dir/a.bam" \
+  --bai "$test_dir/a.bam.bai" \
+  --output "$test_dir/a.flagstat"
+
+echo ">>> Checking whether output exists"
+[ ! -f "$test_dir/a.flagstat" ] && echo "File 'a.flagstat' does not exist!" && exit 1
+
+echo ">>> Checking whether output is non-empty"
+[ ! -s "$test_dir/a.flagstat" ] && echo "File 'a.flagstat' is empty!" && exit 1
+
+echo ">>> Checking whether output is correct"
+diff "$test_dir/a.flagstat" "$test_dir/a_ref.flagstat" || \
+    (echo "Output file a.flagstat does not match expected output" && exit 1)
+
+rm "$test_dir/a.flagstat"
+
+############################################################################################
+
+echo ">>> Testing $meta_functionality_name with singletons in the input"
+
+"$meta_executable" \
+  --bam "$test_dir/test.paired_end.sorted.bam" \
+  --bai "$test_dir/test.paired_end.sorted.bam.bai" \
+  --output "$test_dir/test.paired_end.sorted.flagstat"
+
+echo ">>> Checking whether output exists"
+[ ! -f "$test_dir/test.paired_end.sorted.flagstat" ] && echo "File 'test.paired_end.sorted.flagstat' does not exist!" && exit 1
+
+echo ">>> Checking whether output is non-empty"
+[ ! -s "$test_dir/test.paired_end.sorted.flagstat" ] && echo "File 'test.paired_end.sorted.flagstat' is empty!" && exit 1
+
+echo ">>> Checking whether output is correct"
+diff "$test_dir/test.paired_end.sorted.flagstat" "$test_dir/test_ref.paired_end.sorted.flagstat" || \
+    (echo "Output file test.paired_end.sorted.flagstat does not match expected output" && exit 1)
+
+rm "$test_dir/test.paired_end.sorted.flagstat"
+
+
+
+echo "All tests succeeded!"
+exit 0

src/samtools/samtools_flagstat/test_data/a_ref.flagstat

@@ -0,0 +1,16 @@
+6 + 0 in total (QC-passed reads + QC-failed reads)
+6 + 0 primary
+0 + 0 secondary
+0 + 0 supplementary
+0 + 0 duplicates
+0 + 0 primary duplicates
+6 + 0 mapped (100.00% : N/A)
+6 + 0 primary mapped (100.00% : N/A)
+6 + 0 paired in sequencing
+3 + 0 read1
+3 + 0 read2
+6 + 0 properly paired (100.00% : N/A)
+6 + 0 with itself and mate mapped
+0 + 0 singletons (0.00% : N/A)
+0 + 0 with mate mapped to a different chr
+0 + 0 with mate mapped to a different chr (mapQ>=5)

src/samtools/samtools_flagstat/test_data/script.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+# Download test data from snakemake wrapper
+
+wget https://raw.githubusercontent.com/snakemake/snakemake-wrappers/3a4f7004281efc176fd9af732ad88d00c47d432d/bio/samtools/flagstat/test/mapped/a.bam
+samtools index a.bam
+# samtools flagstat a.bam > a_ref.flagstat
+
+
+# Download test data from nf-core module
+
+wget https://github.com/nf-core/test-datasets/raw/modules/data/genomics/sarscov2/illumina/bam/test.paired_end.sorted.bam
+wget https://github.com/nf-core/test-datasets/raw/modules/data/genomics/sarscov2/illumina/bam/test.paired_end.sorted.bam.bai
+# samtools flagstat test.paired_end.sorted.bam > test_ref.paired_end.sorted.flagstat

src/samtools/samtools_flagstat/test_data/test_ref.paired_end.sorted.flagstat

@@ -0,0 +1,16 @@
+200 + 0 in total (QC-passed reads + QC-failed reads)
+200 + 0 primary
+0 + 0 secondary
+0 + 0 supplementary
+0 + 0 duplicates
+0 + 0 primary duplicates
+197 + 0 mapped (98.50% : N/A)
+197 + 0 primary mapped (98.50% : N/A)
+200 + 0 paired in sequencing
+100 + 0 read1
+100 + 0 read2
+192 + 0 properly paired (96.00% : N/A)
+194 + 0 with itself and mate mapped
+3 + 0 singletons (1.50% : N/A)
+0 + 0 with mate mapped to a different chr
+0 + 0 with mate mapped to a different chr (mapQ>=5)

src/samtools/samtools_idxstats/config.vsh.yaml

@@ -0,0 +1,53 @@
+name: samtools_idxstats
+namespace: samtools
+description: Reports alignment summary statistics for a BAM file.
+keywords: [stats, mapping, counts, chromosome, bam, sam, cram]
+links:
+  homepage: https://www.htslib.org/
+  documentation: https://www.htslib.org/doc/samtools-idxstats.html
+  repository: https://github.com/samtools/samtools
+references: 
+  doi: 10.1093/bioinformatics/btp352, 10.1093/gigascience/giab008
+license: MIT/Expat
+
+argument_groups:
+  - name: Inputs
+    arguments:
+      - name: "--bam"
+        type: file
+        description: BAM input file.
+      - name: "--bai"
+        type: file
+        description: BAM index file.
+      - name: "--fasta"
+        type: file
+        description: Reference file the CRAM was created with (optional).
+  - name: Outputs
+    arguments:
+      - name: "--output"
+        type: file
+        description: |
+          File containing samtools stats output in tab-delimited format.
+        required: true
+        must_exist: false
+        example: output.idxstats
+
+resources:
+  - type: bash_script
+    path: script.sh
+test_resources:
+  - type: bash_script
+    path: test.sh
+  - type: file
+    path: test_data
+engines:
+  - type: docker
+    image: quay.io/biocontainers/samtools:1.19.2--h50ea8bc_1
+    setup:
+      - type: docker
+        run: |
+          samtools --version 2>&1 | grep -E '^(samtools|Using htslib)' | \
+          sed 's#Using ##;s# \([0-9\.]*\)$#: \1#' > /var/software_versions.txt
+runners:
+- type: executable
+- type: nextflow

src/samtools/samtools_idxstats/help.txt

@@ -0,0 +1,12 @@
+```
+samtools idxstats
+```
+
+Usage: samtools idxstats [options] <in.bam>
+      --input-fmt-option OPT[=VAL]
+               Specify a single input file format option in the form
+               of OPTION or OPTION=VALUE
+  -@, --threads INT
+               Number of additional threads to use [0]
+      --verbosity INT
+               Set level of verbosity

src/samtools/samtools_idxstats/script.sh

@@ -0,0 +1,8 @@
+#!/bin/bash
+
+## VIASH START
+## VIASH END
+
+set -e
+
+samtools idxstats "$par_bam" > "$par_output"

src/samtools/samtools_idxstats/test.sh

@@ -0,0 +1,49 @@
+#!/bin/bash
+
+test_dir="${meta_resources_dir}/test_data"
+echo ">>> Testing $meta_functionality_name"
+
+"$meta_executable" \
+  --bam "$test_dir/a.sorted.bam" \
+  --bai "$test_dir/a.sorted.bam.bai" \
+  --output "$test_dir/a.sorted.idxstats"
+
+echo ">>> Checking whether output exists"
+[ ! -f "$test_dir/a.sorted.idxstats" ] && echo "File 'a.sorted.idxstats' does not exist!" && exit 1
+
+echo ">>> Checking whether output is non-empty"
+[ ! -s "$test_dir/a.sorted.idxstats" ] && echo "File 'a.sorted.idxstats' is empty!" && exit 1
+
+echo ">>> Checking whether output is correct"
+diff "$test_dir/a.sorted.idxstats" "$test_dir/a_ref.sorted.idxstats" || \
+    (echo "Output file a.sorted.idxstats does not match expected output" && exit 1)
+
+rm "$test_dir/a.sorted.idxstats"
+
+############################################################################################
+
+echo ">>> Testing $meta_functionality_name with singletons in the input"
+
+"$meta_executable" \
+  --bam "$test_dir/test.paired_end.sorted.bam" \
+  --bai "$test_dir/test.paired_end.sorted.bam.bai" \
+  --output "$test_dir/test.paired_end.sorted.idxstats"
+
+echo ">>> Checking whether output exists"
+[ ! -f "$test_dir/test.paired_end.sorted.idxstats" ] && \
+    echo "File 'test.paired_end.sorted.idxstats' does not exist!" && exit 1
+
+echo ">>> Checking whether output is non-empty"
+[ ! -s "$test_dir/test.paired_end.sorted.idxstats" ] && \
+    echo "File 'test.paired_end.sorted.idxstats' is empty!" && exit 1
+
+echo ">>> Checking whether output is correct"
+diff "$test_dir/test.paired_end.sorted.idxstats" "$test_dir/test_ref.paired_end.sorted.idxstats" || \
+    (echo "Output file test.paired_end.sorted.idxstats does not match expected output" && exit 1)
+
+rm "$test_dir/test.paired_end.sorted.idxstats"
+
+############################################################################################
+
+echo "All tests succeeded!"
+exit 0

src/samtools/samtools_idxstats/test_data/a_ref.sorted.idxstats

@@ -0,0 +1,2 @@
+xx	20	6	0
+*	0	0	0

src/samtools/samtools_idxstats/test_data/script.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+# dowload test data from snakemake wrapper
+if [ ! -d /tmp/idxstats_source ]; then
+  git clone --depth 1 --single-branch --branch master https://github.com/snakemake/snakemake-wrappers.git /tmp/idxstats_source
+fi
+
+cp -r /tmp/idxstats_source/bio/samtools/idxstats/test/mapped/* src/samtools/idxstats/test_data
+# samtools idxstats a.sorted.bam > a.sorted.idxstats
+
+# dowload test data from nf-core module
+wget https://github.com/nf-core/test-datasets/raw/modules/data/genomics/sarscov2/illumina/bam/test.paired_end.sorted.bam
+wget https://github.com/nf-core/test-datasets/raw/modules/data/genomics/sarscov2/illumina/bam/test.paired_end.sorted.bam.bai
+# samtools idxstats test.paired_end.sorted.bam > test_ref.paired_end.sorted.idxstats

src/samtools/samtools_idxstats/test_data/test_ref.paired_end.sorted.idxstats

@@ -0,0 +1,2 @@
+MT192765.1	29829	197	3
+*	0	0	0