Merge pull request #242 from JoseEspinosa/i240

Separate mmcif_files from obsolete channels

CHANGELOG.md

@@ -16,6 +16,17 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [[PR #210](https://github.com/nf-core/proteinfold/pull/210)] - Moving post-processing logic to a subworkflow, change wave images pointing to oras to point to https and refactor module to match nf-core folder structure.
 - [[#214](https://github.com/nf-core/proteinfold/issues/214)] - Fix colabfold image to run in cpus after [#188](https://github.com/nf-core/proteinfold/issues/188) fix.
 - [[#235](https://github.com/nf-core/proteinfold/issues/235)] - Update samplesheet to new version (switch from `sequence` column to `id`).
+- [[#240](https://github.com/nf-core/proteinfold/issues/240)] - Separate download and input of pdb `mmcif` files and `obsolete` database.
+
+### Parameters
+
+| Old parameter | New parameter         |
+| ------------- | --------------------- |
+|               | `--pdb_obsolete_path` |
+
+> **NB:** Parameter has been **updated** if both old and new parameter information is present.
+> **NB:** Parameter has been **added** if just the new parameter information is present.
+> **NB:** Parameter has been **removed** if parameter information is
 
 ## [[1.1.1](https://github.com/nf-core/proteinfold/releases/tag/1.1.1)] - 2025-07-30
 

conf/dbs.config

@@ -29,6 +29,7 @@ params {
     mgnify_path              = "${params.alphafold2_db}/mgnify/*"
     pdb70_path               = "${params.alphafold2_db}/pdb70/**"
     pdb_mmcif_path           = "${params.alphafold2_db}/pdb_mmcif/*"
+    pdb_obsolete_path        = "${params.alphafold2_db}/pdb_mmcif/obsolete.dat"
     uniref30_alphafold2_path = "${params.alphafold2_db}/uniref30/*"
     uniref90_path            = "${params.alphafold2_db}/uniref90/*"
     pdb_seqres_path          = "${params.alphafold2_db}/pdb_seqres/*"

main.nf

@@ -88,6 +88,7 @@ workflow NFCORE_PROTEINFOLD {
             params.mgnify_path,
             params.pdb70_path,
             params.pdb_mmcif_path,
+            params.pdb_obsolete_path,
             params.uniref30_alphafold2_path,
             params.uniref90_path,
             params.pdb_seqres_path,
@@ -122,6 +123,7 @@ workflow NFCORE_PROTEINFOLD {
             PREPARE_ALPHAFOLD2_DBS.out.mgnify,
             PREPARE_ALPHAFOLD2_DBS.out.pdb70,
             PREPARE_ALPHAFOLD2_DBS.out.pdb_mmcif,
+            PREPARE_ALPHAFOLD2_DBS.out.pdb_obsolete,
             PREPARE_ALPHAFOLD2_DBS.out.uniref30,
             PREPARE_ALPHAFOLD2_DBS.out.uniref90,
             PREPARE_ALPHAFOLD2_DBS.out.pdb_seqres,

modules/local/download_pdbmmcif/main.nf

@@ -13,7 +13,6 @@ process DOWNLOAD_PDBMMCIF {
 
     input:
     val source_url_pdb_mmcif
-    val source_url_pdb_obsolete
 
     output:
     path ('*')         , emit: ch_db
@@ -54,14 +53,10 @@ process DOWNLOAD_PDBMMCIF {
     # Delete empty download directory structure.
     find ./raw -type d -empty -delete
 
-    aria2c \\
-        $source_url_pdb_obsolete
-
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
         sed: \$(echo \$(sed --version 2>&1) | head -1 | sed 's/^.*GNU sed) //; s/ .*\$//')
         rsync: \$(rsync --version | head -1 | sed 's/^rsync  version //; s/  protocol version [[:digit:]]*//')
-        aria2c: \$( aria2c -v | head -1 | sed 's/aria2 version //' )
     END_VERSIONS
     """
 

modules/local/run_alphafold2/main.nf

@@ -21,7 +21,8 @@ process RUN_ALPHAFOLD2 {
     path ('small_bfd/*')
     path ('mgnify/*')
     path ('pdb70/*')
-    path ('pdb_mmcif/*')
+    path ('mmcif_files/*')
+    path ('obsolete_pdb/*')
     path ('uniref30/*')
     path ('uniref90/*')
     path ('pdb_seqres/*')
@@ -61,8 +62,8 @@ process RUN_ALPHAFOLD2 {
         --data_dir=\$PWD \
         --uniref90_database_path=./uniref90/uniref90.fasta \
         --mgnify_database_path=./mgnify/mgy_clusters_2022_05.fa \
-        --template_mmcif_dir=./pdb_mmcif/mmcif_files \
-        --obsolete_pdbs_path=./pdb_mmcif/obsolete.dat \
+        --template_mmcif_dir=./mmcif_files \
+        --obsolete_pdbs_path=./obsolete_pdb/obsolete.dat \
         --random_seed=53343 \
         $args
 

modules/local/run_alphafold2_msa/main.nf

@@ -21,6 +21,7 @@ process RUN_ALPHAFOLD2_MSA {
     path ('small_bfd/*')
     path ('mgnify/*')
     path ('pdb70/*')
+    path ('pdb_mmcif/mmcif_files/*')
     path ('pdb_mmcif/*')
     path ('uniref30/*')
     path ('uniref90/*')

modules/local/run_alphafold2_pred/main.nf

@@ -21,6 +21,7 @@ process RUN_ALPHAFOLD2_PRED {
     path ('small_bfd/*')
     path ('mgnify/*')
     path ('pdb70/*')
+    path ('pdb_mmcif/mmcif_files/*')
     path ('pdb_mmcif/*')
     path ('uniref30/*')
     path ('uniref90/*')

nextflow.config

@@ -43,6 +43,7 @@ params {
     mgnify_path                 = null
     pdb70_path                  = null
     pdb_mmcif_path              = null
+    pdb_obsolete_path           = null
     uniref30_alphafold2_path    = null
     uniref90_path               = null
     pdb_seqres_path             = null

nextflow_schema.json

@@ -412,6 +412,11 @@
                     "description": "Path to the PDB mmCIF database",
                     "fa_icon": "fas fa-folder-open"
                 },
+                "pdb_obsolete_path": {
+                    "type": "string",
+                    "description": "Path to the PDB obsolete file",
+                    "fa_icon": "fas fa-folder-open"
+                },
                 "uniref30_alphafold2_path": {
                     "type": "string",
                     "description": "Path to the Uniref30 database",

subworkflows/local/prepare_alphafold2_dbs.nf

@@ -8,6 +8,7 @@ include {
     ARIA2_UNCOMPRESS as ARIA2_SMALL_BFD
     ARIA2_UNCOMPRESS as ARIA2_MGNIFY
     ARIA2_UNCOMPRESS as ARIA2_PDB70
+    ARIA2_UNCOMPRESS as ARIA2_OBSOLETE
     ARIA2_UNCOMPRESS as ARIA2_UNIREF30
     ARIA2_UNCOMPRESS as ARIA2_UNIREF90
     ARIA2_UNCOMPRESS as ARIA2_UNIPROT_SPROT
@@ -28,7 +29,8 @@ workflow PREPARE_ALPHAFOLD2_DBS {
     alphafold2_params_path   // directory: /path/to/alphafold2/params/
     mgnify_path              // directory: /path/to/mgnify/
     pdb70_path               // directory: /path/to/pdb70/
-    pdb_mmcif_path           // directory: /path/to/pdb_mmcif/
+    pdb_mmcif_path           // directory: /path/to/pdb_mmcif/mmcif_files/
+    pdb_obsolete_path        // directory: /path/to/pdb_mmcif/obsolete.dat
     uniref30_alphafold2_path // directory: /path/to/uniref30/alphafold2/
     uniref90_path            // directory: /path/to/uniref90/
     pdb_seqres_path          // directory: /path/to/pdb_seqres/
@@ -65,9 +67,8 @@ workflow PREPARE_ALPHAFOLD2_DBS {
         ch_params         = Channel.value(file(alphafold2_params_path))
         ch_mgnify         = Channel.value(file(mgnify_path))
         ch_pdb70          = Channel.value(file(pdb70_path, type: 'dir' ))
-        ch_mmcif_files    = file(pdb_mmcif_path, type: 'dir')
-        ch_mmcif_obsolete = file(pdb_mmcif_path, type: 'file')
-        ch_mmcif          = Channel.value(ch_mmcif_files + ch_mmcif_obsolete)
+        ch_mmcif_files    = Channel.value(file(pdb_mmcif_path))
+        ch_obsolete       = Channel.value(file(pdb_obsolete_path, type: 'file'))
         ch_uniref30       = Channel.value(file(uniref30_alphafold2_path, type: 'any'))
         ch_uniref90       = Channel.value(file(uniref90_path))
         ch_pdb_seqres     = Channel.value(file(pdb_seqres_path))
@@ -108,10 +109,15 @@ workflow PREPARE_ALPHAFOLD2_DBS {
 
         DOWNLOAD_PDBMMCIF(
             pdb_mmcif_link,
+        )
+        ch_mmcif_files = DOWNLOAD_PDBMMCIF.out.ch_db
+        ch_versions    = ch_versions.mix(DOWNLOAD_PDBMMCIF.out.versions)
+
+        ARIA2_OBSOLETE(
             pdb_obsolete_link
         )
-        ch_mmcif = DOWNLOAD_PDBMMCIF.out.ch_db
-        ch_versions = ch_versions.mix(DOWNLOAD_PDBMMCIF.out.versions)
+        ch_obsolete = ARIA2_OBSOLETE.out.db
+        ch_versions = ch_versions.mix(ARIA2_OBSOLETE.out.versions)
 
         ARIA2_UNIREF30(
             uniref30_alphafold2_link
@@ -146,20 +152,21 @@ workflow PREPARE_ALPHAFOLD2_DBS {
             ARIA2_UNIPROT_SPROT.out.db,
             ARIA2_UNIPROT_TREMBL.out.db
         )
-        ch_uniprot = COMBINE_UNIPROT.out.ch_db
-        ch_version =  ch_versions.mix(COMBINE_UNIPROT.out.versions)
+        ch_uniprot  = COMBINE_UNIPROT.out.ch_db
+        ch_versions = ch_versions.mix(COMBINE_UNIPROT.out.versions)
     }
 
     emit:
-    bfd        = ch_bfd
-    small_bfd  = ch_small_bfd
-    params     = ch_params
-    mgnify     = ch_mgnify
-    pdb70      = ch_pdb70
-    pdb_mmcif  = ch_mmcif
-    uniref30   = ch_uniref30
-    uniref90   = ch_uniref90
-    pdb_seqres = ch_pdb_seqres
-    uniprot    = ch_uniprot
-    versions   = ch_versions
+    bfd          = ch_bfd
+    small_bfd    = ch_small_bfd
+    params       = ch_params
+    mgnify       = ch_mgnify
+    pdb70        = ch_pdb70
+    pdb_mmcif    = ch_mmcif_files
+    pdb_obsolete = ch_obsolete
+    uniref30     = ch_uniref30
+    uniref90     = ch_uniref90
+    pdb_seqres   = ch_pdb_seqres
+    uniprot      = ch_uniprot
+    versions     = ch_versions
 }

workflows/alphafold2.nf

@@ -37,6 +37,7 @@ workflow ALPHAFOLD2 {
     ch_mgnify               // channel: path(mgnify)
     ch_pdb70                // channel: path(pdb70)
     ch_pdb_mmcif            // channel: path(pdb_mmcif)
+    ch_pdb_obsolete         // channel: path(pdb_obsolete)
     ch_uniref30             // channel: path(uniref30)
     ch_uniref90             // channel: path(uniref90)
     ch_pdb_seqres           // channel: path(pdb_seqres)
@@ -73,6 +74,7 @@ workflow ALPHAFOLD2 {
             ch_mgnify,
             ch_pdb70,
             ch_pdb_mmcif,
+            ch_pdb_obsolete,
             ch_uniref30,
             ch_uniref90,
             ch_pdb_seqres,
@@ -106,6 +108,7 @@ workflow ALPHAFOLD2 {
             ch_mgnify,
             ch_pdb70,
             ch_pdb_mmcif,
+            ch_pdb_obsolete,
             ch_uniref30,
             ch_uniref90,
             ch_pdb_seqres,
@@ -123,6 +126,7 @@ workflow ALPHAFOLD2 {
             ch_mgnify,
             ch_pdb70,
             ch_pdb_mmcif,
+            ch_pdb_obsolete,
             ch_uniref30,
             ch_uniref90,
             ch_pdb_seqres,