Add mapping tools and viral pathogen List

.nf-core.yml

@@ -15,6 +15,7 @@ lint:
     - assets/email_template.html
     - assets/email_template.txt
     - docs/README.md
+    - viralPathogenList.csv
 
   multiqc_config:
     - report_comment

CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Extract Kraken2 reads with KrakenTools
 - Extract Centrifuge reads
 - Extract DIAMOND reads
+- Screen pathogens via mapping against genomes from a list of pethogens: Bowtie2 for short reads and Minimap2 for long reads
 
 ### `Fixed`
 

assets/test_data/reference/accession2taxid.map

@@ -0,0 +1,102 @@
+NC_002205.1	518987
+NC_002206.1	518987
+NC_002207.1	518987
+NC_002208.1	518987
+NC_002209.1	518987
+NC_002210.1	518987
+NC_002211.1	518987
+NC_002204.1	518987
+NC_004910.1	130760
+NC_004911.1	130760
+NC_004912.1	130760
+NC_004908.1	130760
+NC_004905.2	130760
+NC_004909.1	130760
+NC_004907.1	130760
+NC_004906.1	130760
+NC_006307.2	11553
+NC_006308.2	11553
+NC_006309.2	11553
+NC_006310.2	11553
+NC_006311.1	11553
+NC_006312.2	11553
+NC_006306.2	11553
+NC_007357.1	93838
+NC_007358.1	93838
+NC_007359.1	93838
+NC_007362.1	93838
+NC_007360.1	93838
+NC_007361.1	93838
+NC_007363.1	93838
+NC_007364.1	93838
+NC_007373.1	335341
+NC_007372.1	335341
+NC_007371.1	335341
+NC_007366.1	335341
+NC_007369.1	335341
+NC_007368.1	335341
+NC_007367.1	335341
+NC_007370.1	335341
+NC_002023.1	211044
+NC_002021.1	211044
+NC_002022.1	211044
+NC_002017.1	211044
+NC_002019.1	211044
+NC_002018.1	211044
+NC_002016.1	211044
+NC_002020.1	211044
+NC_007378.1	488241
+NC_007375.1	488241
+NC_007376.1	488241
+NC_007374.1	488241
+NC_007381.1	488241
+NC_007382.1	488241
+NC_007377.1	488241
+NC_007380.1	488241
+NC_026422.1	1332244
+NC_026423.1	1332244
+NC_026424.1	1332244
+NC_026425.1	1332244
+NC_026426.1	1332244
+NC_026429.1	1332244
+NC_026427.1	1332244
+NC_026428.1	1332244
+NC_026438.1	641809
+NC_026435.1	641809
+NC_026437.1	641809
+NC_026433.1	641809
+NC_026436.1	641809
+NC_026434.1	641809
+NC_026431.1	641809
+NC_026432.1	641809
+NC_036616.1	1173138
+NC_036615.1	1173138
+NC_036619.1	1173138
+NC_036618.1	1173138
+NC_036617.1	1173138
+NC_036620.1	1173138
+NC_036621.1	1173138
+NC_060925.1	9606
+NC_060926.1	9606
+NC_060927.1	9606
+NC_060928.1	9606
+NC_060929.1	9606
+NC_060930.1	9606
+NC_060931.1	9606
+NC_060932.1	9606
+NC_060933.1	9606
+NC_060934.1	9606
+NC_060935.1	9606
+NC_060936.1	9606
+NC_060937.1	9606
+NC_060938.1	9606
+NC_060939.1	9606
+NC_060940.1	9606
+NC_060941.1	9606
+NC_060942.1	9606
+NC_060943.1	9606
+NC_060944.1	9606
+NC_060945.1	9606
+NC_060946.1	9606
+NC_060947.1	9606
+NC_060948.1	9606

conf/modules.config

@@ -24,7 +24,7 @@ process {
         publishDir = [
             path: { "${params.outdir}/fastqc/input" },
             mode: params.publish_dir_mode,
-            pattern: '*.{html,zip}'
+            pattern: '*.{html}'
         ]
     }
 
@@ -36,11 +36,12 @@ process {
             pattern: '*viral_taxids.tsv'
         ]
     }
+
     withName: KRAKENTOOLS_EXTRACTKRAKENREADS {
         ext.args = { params.fastq_output ? "--fastq-output" : ""}
         ext.prefix = { "${meta.id}_${taxid.toString().replaceAll(' ', '-')}" }
         publishDir = [
-            path: { "${params.outdir}/reads/kraken2" },
+            path: { "${params.outdir}/extracted_reads/kraken2" },
             mode: params.publish_dir_mode,
             pattern: '*.{fastq,fasta}'
         ]
@@ -49,26 +50,84 @@ process {
     withName: EXTRACTCENTRIFUGEREADS {
         ext.prefix = { "${meta.id}" }
         publishDir = [
-            path: { "${params.outdir}/reads/centrifuge" },
+            path: { "${params.outdir}/extracted_reads/centrifuge" },
             mode: params.publish_dir_mode,
             pattern: '*.fastq'
         ]
     }
+
     withName: EXTRACTCDIAMONDREADS {
         ext.prefix = { "${meta.id}" }
         publishDir = [
-            path: { "${params.outdir}/reads/diamond" },
+            path: { "${params.outdir}/extracted_reads/diamond" },
             mode: params.publish_dir_mode,
             pattern: '*.fastq'
         ]
     }
+
+    withName: '.*:FASTQ_ALIGN_BOWTIE2:BAM_SORT_STATS_SAMTOOLS:SAMTOOLS_(SORT|INDEX)' {
+        publishDir = [
+            path: { "$params.outdir/pathogens/bowtie2/align" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: 'SAMTOOLS_SORT' {
+        ext.prefix = { "${meta.id}_aligned_pathogens_genome_sorted"}
+    }
+
+    withName: '.*:FASTQ_ALIGN_BOWTIE2:BAM_SORT_STATS_SAMTOOLS:BAM_STATS_SAMTOOLS:.*' {
+        publishDir = [ enabled: false ]
+    }
+
+    withName: 'BOWTIE2_BUILD_PATHOGEN' {
+        ext.args = '--large-index'
+        publishDir = [
+            path: { "$params.outdir/pathogens/bowtie2/build" },
+            mode: params.publish_dir_mode,
+            pattern: 'bowtie2'
+        ]
+    }
+
+    withName: '.*:FASTQ_ALIGN_BOWTIE2:BOWTIE2_ALIGN' {
+        ext.args = '--no-unal'
+        publishDir = [ enabled: false]
+    }
+
+    withName: '.*:LONGREAD_SCREENPATHOGEN:MINIMAP2_INDEX' {
+        ext.args = '-x map-ont'
+        publishDir = [
+            path: { "$params.outdir/pathogens/minimap2/index" },
+            mode: params.publish_dir_mode,
+            pattern: '*.mmi'
+        ]
+    }
+
+    withName: '.*:LONGREAD_SCREENPATHOGEN:MINIMAP2_ALIGN' {
+        ext.args = '--sam-hit-only'
+        publishDir = [ enabled: false]
+    }
+
+    withName: '.*:LONGREAD_SCREENPATHOGEN:BAM_SORT_STATS_SAMTOOLS:SAMTOOLS_(SORT|INDEX)' {
+        publishDir = [
+            path: { "$params.outdir/pathogens/minimap2/align" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: '.*:LONGREAD_SCREENPATHOGEN:BAM_SORT_STATS_SAMTOOLS:BAM_STATS_SAMTOOLS:.*' {
+        publishDir = [ enabled: false ]
+    }
+
     withName: MULTIQC {
         ext.args = { { params.multiqc_title ? "--title \"$params.multiqc_title\"" : '' } }
         publishDir = [
             path: { "${params.outdir}/multiqc"},
             mode: params.publish_dir_mode,
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
         ]
-
     }
+
 }

conf/test.config

@@ -20,18 +20,19 @@ params {
     max_time                   = '6.h'
 
     // Input data
-    //input  = '../final_test_data/samplesheet_v3.csv'
     input                         = 'assets/samplesheet_v1.csv'
-
-    taxid                         = '211044 11676'         //separated by space
+    pathogens_genome              = 'assets/test_data/reference/reference.fna'
 
     // Extract reads
+    perform_extract_reads         = true
     extract_kraken2_reads         = true
     fastq_output                  = true
-
     extract_centrifuge_reads      = true
     extract_diamond_reads         = true
 
+    // Screen pathogens
+    perform_screen_pathogens      = true
+
     // Genome references
     genome = 'R64-1-1'
 }

conf/test_full.config

@@ -15,9 +15,18 @@ params {
     config_profile_description = 'Full test dataset to check pipeline function'
 
     // Input data for full size test
-    // TODO nf-core: Specify the paths to your full test data ( on nf-core/test-datasets or directly in repositories, e.g. SRA)
-    // TODO nf-core: Give any required params for the test so that command line flags are not needed
-    input = params.pipelines_testdata_base_path + 'viralrecon/samplesheet/samplesheet_full_illumina_amplicon.csv'
+    input                         = '../final_test_data/samplesheet_v3.csv'
+    pathogens_genome              = 'assets/test_data/reference/reference.fna'
+
+    // Extract reads
+    perform_extract_reads         = true
+    extract_kraken2_reads         = true
+    fastq_output                  = true
+    extract_centrifuge_reads      = true
+    extract_diamond_reads         = true
+
+    // Screen pathogens
+    perform_screen_pathogens      = true
 
     // Genome references
     genome = 'R64-1-1'

conf/test_taxid.config

@@ -0,0 +1,39 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Nextflow config file for running minimal tests
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Defines a input taxid list to run a fast and simple pipeline test.
+
+    Use as follows:
+        nextflow run genomic-medicine-sweden/meta-val -profile test_taxid,<docker/singularity> --outdir <OUTDIR>
+
+----------------------------------------------------------------------------------------
+*/
+
+params {
+    config_profile_name        = 'Test user defined taxid profile'
+    config_profile_description = 'Minimal test dataset to check pipeline function'
+
+    // Limit resources so that this can run on GitHub Actions
+    max_cpus                   = 2
+    max_memory                 = '6.GB'
+    max_time                   = '6.h'
+
+    // Input data
+    input                         = 'assets/samplesheet_v1.csv'
+    taxid                         = '211044 11676'         //separated by space
+    pathogens_genome              = 'assets/test_data/reference/reference.fna'
+
+    // Extract reads
+    perform_extract_reads         = true
+    extract_kraken2_reads         = true
+    fastq_output                  = true
+    extract_centrifuge_reads      = true
+    extract_diamond_reads         = true
+
+    // Screen pathogens
+    perform_screen_pathogens      = false
+
+    // Genome references
+    genome = 'R64-1-1'
+}

modules.json

@@ -5,6 +5,16 @@
         "https://github.com/nf-core/modules.git": {
             "modules": {
                 "nf-core": {
+                    "bowtie2/align": {
+                        "branch": "master",
+                        "git_sha": "e4bad511789f16d0df39ee306b2cd50418365048",
+                        "installed_by": ["fastq_align_bowtie2", "modules"]
+                    },
+                    "bowtie2/build": {
+                        "branch": "master",
+                        "git_sha": "1fea64f5132a813ec97c1c6d3a74e0aee7142b6d",
+                        "installed_by": ["modules"]
+                    },
                     "fastqc": {
                         "branch": "master",
                         "git_sha": "285a50500f9e02578d90b3ce6382ea3c30216acd",
@@ -15,15 +25,65 @@
                         "git_sha": "77decb880af3f0ea6d22b2383c3f1ed86aac4aa2",
                         "installed_by": ["modules"]
                     },
+                    "minimap2/align": {
+                        "branch": "master",
+                        "git_sha": "e83b347b3e674de6bb1bb7bdf9b2674768c8e0fe",
+                        "installed_by": ["modules"]
+                    },
+                    "minimap2/index": {
+                        "branch": "master",
+                        "git_sha": "72e277acfd9e61a9f1368eafb4a9e83f5bcaa9f5",
+                        "installed_by": ["modules"]
+                    },
                     "multiqc": {
                         "branch": "master",
                         "git_sha": "b7ebe95761cd389603f9cc0e0dc384c0f663815a",
                         "installed_by": ["modules"]
+                    },
+                    "samtools/flagstat": {
+                        "branch": "master",
+                        "git_sha": "04fbbc7c43cebc0b95d5b126f6d9fe4effa33519",
+                        "installed_by": ["bam_stats_samtools"]
+                    },
+                    "samtools/idxstats": {
+                        "branch": "master",
+                        "git_sha": "04fbbc7c43cebc0b95d5b126f6d9fe4effa33519",
+                        "installed_by": ["bam_stats_samtools"]
+                    },
+                    "samtools/index": {
+                        "branch": "master",
+                        "git_sha": "04fbbc7c43cebc0b95d5b126f6d9fe4effa33519",
+                        "installed_by": ["bam_sort_stats_samtools"]
+                    },
+                    "samtools/sort": {
+                        "branch": "master",
+                        "git_sha": "04fbbc7c43cebc0b95d5b126f6d9fe4effa33519",
+                        "installed_by": ["bam_sort_stats_samtools"]
+                    },
+                    "samtools/stats": {
+                        "branch": "master",
+                        "git_sha": "04fbbc7c43cebc0b95d5b126f6d9fe4effa33519",
+                        "installed_by": ["bam_stats_samtools"]
                     }
                 }
             },
             "subworkflows": {
                 "nf-core": {
+                    "bam_sort_stats_samtools": {
+                        "branch": "master",
+                        "git_sha": "0eacd714effe5aac1c1de26593873960b3346cab",
+                        "installed_by": ["fastq_align_bowtie2"]
+                    },
+                    "bam_stats_samtools": {
+                        "branch": "master",
+                        "git_sha": "0eacd714effe5aac1c1de26593873960b3346cab",
+                        "installed_by": ["bam_sort_stats_samtools"]
+                    },
+                    "fastq_align_bowtie2": {
+                        "branch": "master",
+                        "git_sha": "0eacd714effe5aac1c1de26593873960b3346cab",
+                        "installed_by": ["subworkflows"]
+                    },
                     "utils_nextflow_pipeline": {
                         "branch": "master",
                         "git_sha": "5caf7640a9ef1d18d765d55339be751bb0969dfa",