Merge pull request #46 from mboudet/dev_version

Manage multiple annotation versions
gogepp · Oct 20, 2023 · f575f82 · f575f82
2 parents 3a3c44e + cb046a2
commit f575f82
Show file tree

Hide file tree

Showing 67 changed files with 2,610 additions and 6,093 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -5,7 +5,7 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html)
 
-## [0.4.8] Unreleased
+## [0.4.8] 2023-10-20
 
 ### Added
 
@@ -18,6 +18,17 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
   - 'query_param' : optional get parameter to use for the query
   - 'field_param': optional get parameter to use to restrict the results to the gene_field value
   - 'count_param': optional get parameter to restrict the number of results
+- Multiple annotations for the same genome
+  - When adding an annotation, you must now set the '--annot' to set the annotation name.
+  - When integrating data afterward, you can use the --annot tag to specify the annotation you are aiming for.
+  - If you have multiple genes with the same ID, and do not specify '--annot', the results may be variables
+  - You can specify --annotations multiple time when integrating orthogroups
+- Will now decode proteins and genes IDs when integrating data. (It was already done when integrating gffs, so there was some mismatch with IDs)
+
+## Changed
+
+- Various UI fixes to better fit multiple annotation versions
+  - Including an 'annotation' selector in the gene list
 
 ## [0.4.7] 2023-09-26
 

diff --git a/cli/genoboo.js b/cli/genoboo.js
@@ -318,10 +318,14 @@ addAnnotation
   .arguments('<file>')
   .option('-u, --username <username>', 'GeneNoteBook admin username')
   .option('-p, --password <password>', 'GeneNoteBook admin password')
-  .option(
+  .requiredOption(
     '-n, --name <genome-name>',
     'Reference genome name to which the annotation should be added.',
   )
+  .requiredOption(
+    '--annot <annotation-name>',
+    'Annotation name',
+  )
   .option(
     '-r, --re_protein <pattern>',
     'Replacement string for the protein name using capturing groups defined by --re_protein_capture. Make sure to use JS-style groups ($1 for group 1)',
@@ -352,6 +356,7 @@ addAnnotation
         password,
         port = 3000,
         name,
+        annot,
         re_protein,
         re_protein_capture,
         attr_protein,
@@ -389,6 +394,7 @@ addAnnotation
         {
           fileName,
           genomeName: name,
+          annotationName: annot,
           re_protein,
           re_protein_capture: correctProteinCapture,
           attr_protein,
@@ -433,6 +439,10 @@ running GeneNoteBook server.`
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '-fmt, --format [parser]',
     `Choose a parser for the diamond output format. Parses .xml, .txt
@@ -455,7 +465,7 @@ running GeneNoteBook server.`
   .action(
     (
       file,
-      { username, password, port = 3000, format, algorithm, matrix, database }
+      { username, password, port = 3000, annot, format, algorithm, matrix, database }
     ) => {
       if (typeof file !== 'string') addDiamond.help();
 
@@ -490,6 +500,7 @@ file extension is not "xml", "txt"`);
           fileName,
           parser: parserType,
           program: 'diamond',
+          annot: annot,
           algorithm: algorithm,
           matrix: matrix,
           database: database,
@@ -502,7 +513,7 @@ file extension is not "xml", "txt"`);
 Example:
     genenotebook add diamond mmucedo.xml -u admin -p admin
 or
-    genenotebook add diamond mmucedo.txt --format txt --program blastp --matrix BLOSUM90 -db "Non-reundant protein sequences (nr)" -u admin -p admin
+    genenotebook add diamond mmucedo.txt --format txt --matrix BLOSUM90 -db "Non-reundant protein sequences (nr)" -u admin -p admin
     `);
   })
   .exitOverride(customExitOverride(addDiamond));
@@ -525,6 +536,10 @@ running GeneNoteBook server.`
     '-p, --password <adminPassword>',
     'GeneNoteBook admin password'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
@@ -548,7 +563,7 @@ running GeneNoteBook server.`
   .action(
     (
       file,
-      { username, password, port = 3000, format, algorithm, matrix, database }
+      { username, password, port = 3000, annot, format, algorithm, matrix, database }
     ) => {
       if (typeof file !== 'string') addBlast.help();
 
@@ -584,6 +599,7 @@ file extension is not "xml", "txt"`);
           parser: parserType,
           program: 'blast',
           algorithm: algorithm,
+          annot: annot,
           matrix: matrix,
           database: database,
         }
@@ -619,6 +635,10 @@ addExpression
     '-d, --sample-description <description>',
     'Description of the experiment'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '-r, --replicas <replicas...>',
     'Comma-separated column positions, which are part of the same replica group. Can be set multiple times for multiple groups. The replica group name will be the first column, unless replica-names is set'
@@ -639,7 +659,7 @@ addExpression
     const replicas = opts.replicas || [];
     const replicaNames = opts.replicaNames || [];
     const isPublic = opts.public;
-
+    const annot = opts.annot
     if (!(fileName && username && password)) {
       program.help();
     }
@@ -648,6 +668,7 @@ addExpression
       {
         fileName,
         description,
+        annot,
         replicas,
         replicaNames,
         isPublic
@@ -680,6 +701,10 @@ addExpression
       '-d, --sample-description <description>',
       'Description of the experiment'
     )
+    .option(
+      '--annot <annotation-ame>',
+      'Annotation name',
+    )
     .option(
       '--public',
       'Set the generated replica groups as public. Default: false',
@@ -692,6 +717,7 @@ addExpression
       const replicaGroup = opts.replicaGroup || fileName;
       const description = opts.sampleDescription || 'description';
       const isPublic = opts.public
+      const anot = opts.annot
 
       if (!(fileName && username && password)) {
         program.help();
@@ -701,6 +727,7 @@ addExpression
         {
           fileName,
           sampleName,
+          annot: annot,
           replicaGroup,
           description,
           isPublic
@@ -728,12 +755,16 @@ addInterproscan
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '--format [parser]',
     `Choose a parser for the interproscan output files. Parses .gff3 and .tsv
     extensions.`
   )
-  .action((file, { username, password, port = 3000, format }) => {
+  .action((file, { username, password, port = 3000, format, annot }) => {
     if (typeof file !== 'string') addInterproscan.help();
 
     const fileName = path.resolve(file);
@@ -766,6 +797,7 @@ file extension is not "tsv", "gff3" or "xml".`);
       {
         fileName,
         parser: parserType,
+        annot: annot
       }
     );
   })
@@ -794,11 +826,15 @@ addEggnog
     '-p, --password <adminPassword>',
     'GeneNoteBook admin password'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
   )
-  .action((file, { username, password, port = 3000 }) => {
+  .action((file, { username, password, port = 3000, annot }) => {
     if (typeof file !== 'string') addEggnog.help();
 
     const fileName = path.resolve(file);
@@ -808,6 +844,7 @@ addEggnog
 
     new GeneNoteBookConnection({ username, password, port }).call('addEggnog', {
       fileName,
+      annot: annot
     });
   })
   .on('--help', () => {
@@ -833,11 +870,15 @@ addHectar
     '-p, --password <adminPassword>',
     'GeneNoteBook admin password'
   )
+  .option(
+    '--annot <annotation-ame>',
+    'Annotation name',
+  )
   .option(
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
   )
-  .action((file, { username, password, port = 3000 }) => {
+  .action((file, { username, password, port = 3000, annot }) => {
     if (typeof file !== 'string') addHectar.help();
 
     const fileName = path.resolve(file);
@@ -847,6 +888,7 @@ addHectar
 
     new GeneNoteBookConnection({ username, password, port }).call('addHectar', {
       fileName,
+      annot: annot
     });
   })
   .on('--help', () => {
@@ -878,17 +920,23 @@ addOrthogroups
     '-f, --force [force]',
     'Ignore the use of prefixes.',
   )
+  .option(
+    '-a, --annotations <annotations...>',
+    'Name of the annotation to use for gene matching. Can be set multiple times'
+  )
   .requiredOption('-u, --username <username>', 'GeneNoteBook admin username')
   .requiredOption('-p, --password <password>', 'GeneNoteBook admin password')
   .option(
     '--port [port]',
     'Port on which GeneNoteBook is running. Default: 3000'
   )
-  .action((file, { prefixe, list, force, username, password, port = 3000 }) => {
+  .action((file, { prefixe, list, force, annotations, username, password, port = 3000 }) => {
     if (typeof file !== 'string') addOrthogroups.help();
     const folderName = path.resolve(file);
     const prefixes = (typeof list !== 'undefined' ? list : path.resolve(prefixe));
 
+    annotations = annotations || []
+
     if (!(folderName && username && password)) {
       addOrthogroups.help();
     }
@@ -899,6 +947,7 @@ addOrthogroups
         folderName,
         force,
         prefixes,
+        annotations
       },
     );
   })

diff --git a/imports/api/api.js b/imports/api/api.js
@@ -17,7 +17,6 @@ import './genomes/removeGenome.js';
 import './genomes/annotation/removeAnnotationTrack.js';
 import './genomes/annotation/addAnnotation.js';
 
-import './genes/interproscan.js';
 import './genes/addInterproscan.js';
 import './genes/eggnog/addEggnog.js';
 import './genes/hectar/addHectar.js';

diff --git a/imports/api/genes/addInterproscan.js b/imports/api/genes/addInterproscan.js
@@ -12,14 +12,16 @@ import { Meteor } from 'meteor/meteor';
  * @method finalize
  */
 class InterproscanProcessor {
-  constructor() {
+  constructor(annot) {
     this.bulkOp = interproscanCollection.rawCollection().initializeUnorderedBulkOp();
     this.geneBulkOp = Genes.rawCollection().initializeUnorderedBulkOp();
     this.currentProt = ""
     this.currentGene = ""
     this.currentContent = []
     this.currentDB = []
-    this.currentOnto = []
+    this.currentOnto = [],
+    this.currentAnnotationName = "",
+    this.annot = annot
   }
 
   finalize = () => {
@@ -31,7 +33,7 @@ class InterproscanProcessor {
     if (this.bulkOp.length > 0){
       return this.bulkOp.execute();
     }
-    return { nMatched: 0 }
+    return { nUpserted: 0 }
   }
 
   updateGenes = () => {
@@ -47,11 +49,13 @@ class InterproscanProcessor {
       this.bulkOp.find({
         gene_id: this.currentGene,
         protein_id: this.currentProt,
+        annotationName: this.currentAnnotationName
       }).upsert().update(
         {
           $set: {
             gene_id: this.currentGene,
             protein_id: this.currentProt,
+            annotationName: this.currentAnnotationName,
             protein_domains: this.currentContent
           },
         },
@@ -63,7 +67,7 @@ class InterproscanProcessor {
     }
 
     if (this.currentDB != [] || this.currentOnto != []){
-      this.geneBulkOp.find({ID: this.currentGene}).update({
+      this.geneBulkOp.find({ID: this.currentGene, annotationName: this.currentAnnotationName}).update({
         $addToSet: {
           'attributes.Ontology_term': { $each: this.currentOnto },
           'attributes.Dbxref': { $each: this.currentDB }
@@ -80,6 +84,10 @@ const addInterproscan = new ValidatedMethod({
   name: 'addInterproscan',
   validate: new SimpleSchema({
     fileName: { type: String },
+    annot: {
+      type: String,
+      optional: true,
+    },
     parser: {
       type: String,
       allowedValues: ['tsv', 'gff3'],
@@ -88,15 +96,15 @@ const addInterproscan = new ValidatedMethod({
   applyOptions: {
     noRetry: true,
   },
-  run({ fileName, parser }) {
+  run({ fileName, annot, parser }) {
     if (!this.userId) {
       throw new Meteor.Error('not-authorized');
     }
     if (!Roles.userIsInRole(this.userId, 'admin')) {
       throw new Meteor.Error('not-authorized');
     }
 
-    const job = new Job(jobQueue, 'addInterproscan', { fileName, parser });
+    const job = new Job(jobQueue, 'addInterproscan', { fileName, annot, parser });
     const jobId = job.priority('high').save();
 
     // Continue with synchronous processing