Merge pull request #178 from daichengxin/diann

restruct DIA-NN step2

modules/local/diann_preliminary_analysis/main.nf

@@ -0,0 +1,63 @@
+process DIANN_PRELIMINARY_ANALYSIS {
+    label 'process_high'
+
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://containers.biocontainers.pro/s3/SingImgsRepo/diann/v1.8.1_cv1/diann_v1.8.1_cv1.img' :
+        'biocontainers/diann:v1.8.1_cv1' }"
+
+    input:
+    tuple file(mzML), file(predict_tsv), file(diann_config)
+
+    output:
+    path "*.quant", emit: diann_quant
+    path "${mzML.baseName}_lib.tsv", emit: lib
+    path "*.log.txt", emit: log
+    path "versions.yml", emit: version
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    min_pr_mz = params.min_pr_mz ? "--min-pr-mz $params.min_pr_mz" : ""
+    max_pr_mz = params.max_pr_mz ? "--max-pr-mz $params.max_pr_mz" : ""
+    min_fr_mz = params.min_fr_mz ? "--min-fr-mz $params.min_fr_mz" : ""
+    max_fr_mz = params.max_fr_mz ? "--max-fr-mz $params.max_fr_mz" : ""
+
+    quick_mass_acc = params.quick_mass_acc ? "--quick-mass-acc" : ""
+    time_corr_only = params.time_corr_only ? "--time-corr-only" : ""
+
+    """
+    diann   `cat diann_config.cfg` \\
+            --lib ${predict_tsv} \\
+            --f ${mzML} \\
+            ${min_pr_mz} \\
+            ${max_pr_mz} \\
+            ${min_fr_mz} \\
+            ${max_fr_mz} \\
+            --threads ${task.cpus} \\
+            --missed-cleavages $params.allowed_missed_cleavages \\
+            --min-pep-len $params.min_peptide_length \\
+            --max-pep-len $params.max_peptide_length \\
+            --min-pr-charge $params.min_precursor_charge \\
+            --max-pr-charge $params.max_precursor_charge \\
+            --var-mods $params.max_mods \\
+            --verbose $params.diann_debug \\
+            --window $params.scan_window \\
+            --gen-spec-lib \\
+            --out-lib ${mzML.baseName}_lib.tsv \\
+            --temp ./ \\
+            --min-corr $params.min_corr \\
+            --corr-diff $params.corr_diff \\
+            ${quick_mass_acc} \\
+            ${time_corr_only} \\
+            $args \\
+            |& tee diann.log
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        DIA-NN: \$(diann 2>&1 | grep "DIA-NN" | grep -oP "(\\d*\\.\\d+\\.\\d+)|(\\d*\\.\\d+)")
+    END_VERSIONS
+    """
+}

modules/local/diann_preliminary_analysis/meta.yml

@@ -0,0 +1,43 @@
+name: diann_preliminary_analysis
+description: A module for preliminary analysis of individual raw files based on DIA-NN.
+keywords:
+  - DIA-NN
+  - DIA
+tools:
+  - DIA-NN:
+      description: |
+        DIA-NN - a universal software for data-independent acquisition (DIA) proteomics data processing by Demichev.
+      homepage: https://github.com/vdemichev/DiaNN
+      documentation: https://github.com/vdemichev/DiaNN
+input:
+  - predict_tsv:
+      type: file
+      description: Silico-predicted spectral library by deep leaning predictor in DIA-NN
+      pattern: "*.tsv"
+  - mzML:
+      type: file
+      description: Spectra file in mzML format
+      pattern: "*.mzML"
+  - cfg:
+      type: dir
+      description: Specifies a file to load options/commands from.
+      pattern: "*.cfg"
+output:
+  - diann_quant:
+      type: file
+      description: Quantification file from DIA-NN
+      pattern: "*.quant"
+  - lib:
+      type: file
+      description: Spectral library file
+      pattern: "*.tsv"
+  - log:
+      type: file
+      description: DIA-NN log file
+      pattern: "diann_report.log.txt"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "versions.yml"
+authors:
+  - "@daichengxin"

nextflow.config

@@ -147,6 +147,12 @@ params {
     // DIA-NN
     matrix_spec_q   = 0.01
     diann_debug     = 3
+    quick_mass_acc  = false
+    scan_window     = 7
+    min_corr        = 2.0
+    corr_diff       = 1.0
+    time_corr_only  = true
+
     // TODO think about unifying it with DDA parameters
     min_pr_mz       = null
     max_pr_mz       = null

nextflow_schema.json

@@ -818,6 +818,38 @@
                     "default": 0.01,
                     "fa_icon": "fas fa-filter"
                 },
+                "quick_mass_acc": {
+                    "type": "boolean",
+                    "description": "Choosing the MS2 mass accuracy setting automatically",
+                    "default": false,
+                    "fa_icon": "fas fa-toggle-on",
+                    "help_text": "(experimental) when choosing the MS2 mass accuracy setting automatically, DIA-NN will use a fast heuristical algorithm instead of IDs number optimisation"
+                },
+                "scan_window": {
+                    "type": "integer",
+                    "description": "Set the scan window radius to a specific value",
+                    "fa_icon": "fas fa-sliders-h",
+                    "help_text": " Ideally, should be approximately equal to the average number of data points per peak",
+                    "default": 7
+                },
+                "min_corr": {
+                    "type": "number",
+                    "description": "Only peaks with correlation sum exceeding min_corr will be considered",
+                    "fa_icon": "fas fa-filter",
+                    "default": 2.0
+                },
+                "corr_diff": {
+                    "type": "number",
+                    "description": "Peaks with correlation sum below corr_diff from maximum will not be considered",
+                    "fa_icon": "fas fa-filter",
+                    "default": 1.0
+                },
+                "time_corr_only": {
+                    "type": "boolean",
+                    "description": "A single score will be used until RT alignment to save memory",
+                    "fa_icon": "fas fa-filter",
+                    "default": true
+                },
                 "min_pr_mz": {
                     "type": "number",
                     "description": "The minimum precursor m/z for the in silico library generation or library-free search",