Merge pull request #284 from ggabernet/fix-reference-data

Use cached reference data by default
nf-core · Oct 23, 2023 · daa1bbd · daa1bbd
2 parents 6845784 + f695dd8
commit daa1bbd
Show file tree

Hide file tree

Showing 8 changed files with 19 additions and 80 deletions.
diff --git a/.github/workflows/ci_immcantation.yml b/.github/workflows/ci_immcantation.yml
@@ -1,9 +1,6 @@
 name: nf-core CI immcantation
 # This workflow runs the pipeline with the minimal test dataset to check that it completes without any syntax errors
 on:
-  push:
-    branches:
-      - dev
   pull_request:
     branches-ignore:
       - "master"

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -225,7 +225,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - [#114](https://github.com/nf-core/bcellmagic/pull/114): Added Bcellmagic html report.
 - [#114](https://github.com/nf-core/bcellmagic/pull/114): Improved documentation on amplicon protocol support.
 - [#115](https://github.com/nf-core/bcellmagic/pull/115): Improved output file structure and documentation.
-- [#124](https://github.com/nf-core/bcellmagic/pull/124): Template update to nf-core tools v2.0.1
+- [#124](https://github.com/nf-core/bcellmagic/pull/124): Template update to nf-core tools v2.0.1.
 
 ### `Fixed`
 
@@ -236,6 +236,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - [#104](https://github.com/nf-core/bcellmagic/pull/104): Fix bug in pairseq barcode copy before consensus.
 - [#114](https://github.com/nf-core/bcellmagic/pull/114): Analysis not restricted to Ig heavy chains.
 - [#123](https://github.com/nf-core/bcellmagic/pull/123): Fix report Rmarkdown reading for running on AWS.
+- [#284](https://github.com/nf-core/bcellmagic/pull/284): Use cached IMGT and IgBlast reference data by default.
 
 ### `Dependencies`
 

diff --git a/conf/test_fetchimgt.config b/conf/test_fetchimgt.config
@@ -23,6 +23,7 @@ params {
     input = 'https://raw.githubusercontent.com/nf-core/test-datasets/airrflow/testdata-bcr/Metadata_test_airr.tsv'
     cprimers = 'https://raw.githubusercontent.com/nf-core/test-datasets/airrflow/testdata-bcr/C_primers.fasta'
     vprimers = 'https://raw.githubusercontent.com/nf-core/test-datasets/airrflow/testdata-bcr/V_primers.fasta'
+    fetch_imgt = true
 
     mode = 'fastq'
 

diff --git a/main.nf b/main.nf
@@ -30,7 +30,7 @@ if (params.help) {
 
 // Validate input parameters
 if (params.validate_params) {
-   validateParameters()
+    validateParameters()
 }
 
 WorkflowMain.initialise(workflow, params, log)

diff --git a/modules/local/igblast/igblast_assigngenes.nf b/modules/local/igblast/igblast_assigngenes.nf
diff --git a/nextflow.config b/nextflow.config
@@ -60,8 +60,9 @@ params {
     // -----------------------
     productive_only = true
     reassign = true
-    igblast_base = null
-    imgtdb_base = null
+    igblast_base = 'https://raw.githubusercontent.com/nf-core/test-datasets/airrflow/database-cache/igblast_base.zip'
+    imgtdb_base = 'https://raw.githubusercontent.com/nf-core/test-datasets/airrflow/database-cache/imgtdb_base.zip'
+    fetch_imgt = false
     save_databases = true
 
     // -----------------------

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -45,7 +45,7 @@
                 },
                 "miairr": {
                     "type": "string",
-                    "default": "bcellmagic/assets/reveal/mapping_MiAIRR_BioSample_v1.3.1.tsv",
+                    "default": "airrflow/assets/reveal/mapping_MiAIRR_BioSample_v1.3.1.tsv",
                     "description": "Path to MiAIRR-BioSample mapping",
                     "fa_icon": "fas fa-table"
                 }
@@ -286,14 +286,19 @@
                 "imgtdb_base": {
                     "type": "string",
                     "description": "Path to the cached IMGT database.",
-                    "help_text": "If it is not provided, the database will be newly downloaded.",
+                    "help_text": "By default, we provide a pre-downloaded version of the IMGT database. It is also possible to provide a custom IMGT reference database. To fetch a fresh version of IMGT, set the `--fetch_imgt` parameter instead.",
                     "fa_icon": "fas fa-database"
                 },
                 "igblast_base": {
                     "type": "string",
                     "description": "Path to the cached igblast database.",
-                    "help_text": "If it is not provided, the database will be newly downloaded.",
+                    "help_text": "By default, we provide a pre-downloaded version of the IMGT database. It is also possible to provide a custom IMGT reference database. To fetch a fresh version of IMGT, set the `--fetch_imgt` parameter instead.",
                     "fa_icon": "fas fa-database"
+                },
+                "fetch_imgt": {
+                    "type": "boolean",
+                    "description": "Set this flag to fetch the IMGT reference data at runtime.",
+                    "fa_icon": "fas fa-cloud-download-alt"
                 }
             },
             "fa_icon": "fas fa-edit"

diff --git a/subworkflows/local/vdj_annotation.nf b/subworkflows/local/vdj_annotation.nf
@@ -4,7 +4,6 @@ include { UNZIP_DB as UNZIP_IMGT } from '../../modules/local/unzip_db'
 include { CHANGEO_ASSIGNGENES } from '../../modules/local/changeo/changeo_assigngenes'
 include { CHANGEO_MAKEDB } from '../../modules/local/changeo/changeo_makedb'
 include { CHANGEO_PARSEDB_SPLIT } from '../../modules/local/changeo/changeo_parsedb_split'
-include { IGBLAST_ASSIGNGENES } from '../../modules/local/igblast/igblast_assigngenes'
 // reveal
 include { FILTER_QUALITY  } from '../../modules/local/reveal/filter_quality'
 include { FILTER_JUNCTION_MOD3  } from '../../modules/local/reveal/filter_junction_mod3'
@@ -25,7 +24,7 @@ workflow VDJ_ANNOTATION {
     // TODO: this can take a long time, and the progress shows 0%. Would be
     // nice to have some better progress reporting.
     // And maybe run this as 2 separate steps, one for IMGT and one for IgBLAST?
-    if( params.igblast_base ){
+    if( !params.fetch_imgt ){
         if (params.igblast_base.endsWith(".zip")) {
             Channel.fromPath("${params.igblast_base}")
                     .ifEmpty{ error "IGBLAST DB not found: ${params.igblast_base}" }
@@ -40,7 +39,7 @@ workflow VDJ_ANNOTATION {
         }
     }
 
-    if( params.imgtdb_base ){
+    if( !params.fetch_imgt ){
         if (params.imgtdb_base.endsWith(".zip")) {
             Channel.fromPath("${params.imgtdb_base}")
                     .ifEmpty{ error "IMGTDB not found: ${params.imgtdb_base}" }
@@ -50,12 +49,12 @@ workflow VDJ_ANNOTATION {
             ch_versions = ch_versions.mix(UNZIP_IMGT.out.versions.ifEmpty(null))
         } else {
             Channel.fromPath("${params.imgtdb_base}")
-                .ifEmpty { error "IMGTDB not found: ${params.imgtdb_base}" }
+                .ifEmpty { error "IMGT DB not found: ${params.imgtdb_base}" }
                 .set { ch_imgt }
         }
     }
 
-    if (!params.igblast_base | !params.imgtdb_base) {
+    if (params.fetch_imgt) {
         FETCH_DATABASES()
         ch_igblast = FETCH_DATABASES.out.igblast
         ch_imgt = FETCH_DATABASES.out.imgt