5.0.6.3 +lis-map and other tweaks

legumeinfo · Apr 26, 2022 · f10f31f · f10f31f
1 parent 6fb254a
commit f10f31f
Show file tree

Hide file tree

Showing 16 changed files with 504 additions and 302 deletions.
diff --git a/build.gradle b/build.gradle
@@ -21,7 +21,7 @@ ext {
 
 subprojects {
     group = 'org.intermine'
-    version = '5.0.6.2'
+    version = '5.0.6.3'
 
     apply plugin: 'java'
     apply plugin: 'maven'

diff --git a/gradle.properties b/gradle.properties
@@ -1,2 +1,2 @@
-systemProp.imVersion=5.0.6.2
-systemProp.bioVersion=5.0.6.2
+systemProp.imVersion=5.0.6.3
+systemProp.bioVersion=5.0.6.3
diff --git a/lis-annotation/src/main/java/org/intermine/bio/dataconversion/AnnotationFileConverter.java b/lis-annotation/src/main/java/org/intermine/bio/dataconversion/AnnotationFileConverter.java
@@ -458,6 +458,7 @@ void processGFF3File() throws IOException, RuntimeException {
             String dbxref = featureI.getAttribute("Dbxref");
             String ontology_term = featureI.getAttribute("Ontology_term");
             String alleles = featureI.getAttribute("alleles");
+            String symbol = featureI.getAttribute("symbol");
             // check that id exists and matches collection
             if (id==null) {
                 throw new RuntimeException("GFF line does not include ID: "+featureI.toString());
@@ -487,9 +488,11 @@ void processGFF3File() throws IOException, RuntimeException {
                 feature.setAttribute("length", String.valueOf(location.length()));
             }
             // Name=GlymaLee.02G198600;
-            if (name!=null) {
-                feature.setAttribute("name", name);
-            }
+            if (name!=null) feature.setAttribute("name", name);
+            // Note=Cytochrome P450 superfamily protein%3B IPR001128 (Cytochrome P450)%3B GO:0005506 (iron ion binding)%2C GO:0020037 (heme binding)%2C ...
+            if (note!=null) feature.setAttribute("description", note);
+            // Symbol=RGB4
+            if (symbol!=null) feature.setAttribute("symbol", symbol);
             // Dbxref=Gene3D:G3DSA:1.10.630.10,InterPro:IPR001128,InterPro:IPR002401,InterPro:IPR017972,PANTHER:PTHR24298,...
             if (dbxref!=null) {
                 String[] terms = dbxref.split(",");
@@ -508,10 +511,6 @@ void processGFF3File() throws IOException, RuntimeException {
                     }
                 }
             }
-            // Note=Cytochrome P450 superfamily protein%3B IPR001128 (Cytochrome P450)%3B GO:0005506 (iron ion binding)%2C GO:0020037 (heme binding)%2C ...
-            if (note!=null) {
-                feature.setAttribute("description", note);
-            }
             // Parent is not required but must already be loaded if present
             if (parent!=null) {
                 Item parentItem = features.get(parent);

diff --git a/lis-datastore/libs/ncgr-datastore.jar b/lis-datastore/libs/ncgr-datastore.jar
diff --git a/lis-genetic/src/main/java/org/intermine/bio/dataconversion/GeneticFileConverter.java b/lis-genetic/src/main/java/org/intermine/bio/dataconversion/GeneticFileConverter.java
diff --git a/lis-genetic/src/main/resources/lis-genetic_additions.xml b/lis-genetic/src/main/resources/lis-genetic_additions.xml
@@ -1,67 +1,40 @@
 <?xml version="1.0"?>
 <classes>
 
-  <!-- extends Annotatable: primaryIdentifier, ontologyAnnotations, publications -->
-  <!-- NOTE: some maps do not have linkage groups, so we must have GeneticMap.qtls as well -->
-  <class name="GeneticMap" extends="Annotatable" is-interface="true">
+  <!-- a LinkageGroup is a component of a GeneticMap, loaded in lis-map, keyed by identifier -->
+  <class name="LinkageGroup" extends="Annotatable" is-interface="true" term="http://purl.obolibrary.org/obo/SO:0000018">
+    <collection name="qtls" referenced-type="QTL" reverse-reference="linkageGroup"/>
+  </class>
+
+  <!-- a QTLStudy represents a particular QTL study from a DataSet/Publication -->
+  <class name="QTLStudy" extends="Annotatable" is-interface="true">
     <attribute name="synopsis" type="java.lang.String"/>
     <attribute name="description" type="java.lang.String"/>
     <attribute name="genotypes" type="java.lang.String"/>
-    <attribute name="genotypingPlatform" type="java.lang.String"/>
-    <attribute name="genotypingMethod" type="java.lang.String"/>
+    <attribute name="geneticMap" type="java.lang.String"/>
     <reference name="organism" referenced-type="Organism"/>
-    <collection name="linkageGroups" referenced-type="LinkageGroup" reverse-reference="geneticMap"/>
-    <collection name="qtls" referenced-type="QTL" reverse-reference="geneticMap"/>
-    <collection name="dataSets" referenced-type="DataSet"/>
-  </class>
-
-  <!-- a linkage group is a part of a genetic map and is related to markers and QTLs -->
-  <!-- marker names are stored in a |-delimited markerNames string; markers are loaded by a post-processor -->
-  <!-- marker reverse reference is via LinkageGroupPosition-->
-  <class name="LinkageGroup" is-interface="true" term="http://purl.obolibrary.org/obo/SO:0000018">
-    <attribute name="identifier" type="java.lang.String"/>
-    <attribute name="number" type="java.lang.Integer"/>
-    <attribute name="length" type="java.lang.Double"/>
-    <attribute name="markerNames" type="java.lang.String"/>
-    <reference name="geneticMap" referenced-type="GeneticMap" reverse-reference="linkageGroups"/>
-    <collection name="markers" referenced-type="GeneticMarker"/>
-    <collection name="qtls" referenced-type="QTL" reverse-reference="linkageGroup"/>
-    <collection name="dataSets" referenced-type="DataSet"/>
-  </class>
-
-  <!-- a genetic marker is a sequence feature mapped to a genome assembly; these gives its cM position on linkage groups -->
-  <class name="GeneticMarker" extends="SequenceFeature" is-interface="true" term="http://purl.obolibrary.org/obo/SO:0001645" >
-    <collection name="linkageGroupPositions" referenced-type="LinkageGroupPosition"/>
-    <collection name="qtls" referenced-type="QTL" reverse-reference="markers"/>
-    <collection name="gwasResults" referenced-type="GWASResult" reverse-reference="markers"/>
-  </class>
-
-  <!-- a marker's position (in cM) on a linkage group, referenced by name -->
-  <class name="LinkageGroupPosition" is-interface="true">
-    <attribute name="markerName" type="java.lang.String"/>
-    <attribute name="position" type="java.lang.Double"/>
-    <reference name="linkageGroup" referenced-type="LinkageGroup"/>
+    <reference name="dataSet" referenced-type="DataSet"/>
+    <collection name="qtls" referenced-type="QTL" reverse-reference="qtlStudy"/>
   </class>
 
-  <!-- QTL references a trait, resides on a linkage group (on a genetic map), and has associated markers -->
-  <!-- marker names are stored in a |-delimited markerNames string; markers are loaded by a post-processor -->
-  <!-- NOTE 1: some maps do not have linkage groups, so we must have QTL.geneticMap reference here -->
-  <!-- NOTE 2: genes are populated by a post-processor -->
-  <class name="QTL" is-interface="true" term="http://purl.obolibrary.org/obo/SO:0001645">
-    <attribute name="identifier" type="java.lang.String"/>
+  <!-- QTL references a trait, resides on a linkage group, and has associated markers -->
+  <!-- marker names are stored here in a |-delimited markerNames string; -->
+  <!-- QTL.markers are loaded by a post-processor from markerNames -->
+  <!-- QTL.genes are populated by a post-processor -->
+  <class name="QTL" extends="Annotatable" is-interface="true" term="http://purl.obolibrary.org/obo/SO:0001645">
     <attribute name="lod" type="java.lang.Double"/>
     <attribute name="likelihoodRatio" type="java.lang.Double"/>
     <attribute name="markerR2" type="java.lang.Double"/>
     <attribute name="peak" type="java.lang.Double"/>
     <attribute name="start" type="java.lang.Double"/>
     <attribute name="end" type="java.lang.Double"/>
     <attribute name="markerNames" type="java.lang.String"/>
+    <reference name="qtlStudy" referenced-type="QTLStudy" reverse-reference="qtls"/>
     <reference name="trait" referenced-type="Trait" reverse-reference="qtls"/>
-    <reference name="linkageGroup" referenced-type="LinkageGroup" reverse-reference="qtls"/>
-    <reference name="geneticMap" referenced-type="GeneticMap" reverse-reference="qtls"/>
     <collection name="markers" referenced-type="GeneticMarker" reverse-reference="qtls"/>
-    <collection name="genes" referenced-type="Gene" reverse-reference="qtls"/>
-    <collection name="dataSets" referenced-type="DataSet"/>
+    <reference name="linkageGroup" referenced-type="LinkageGroup" reverse-reference="qtls"/>
+    <reference name="dataSet" referenced-type="DataSet"/>
+    <collection name="genes" referenced-type="Gene"/>
   </class>
 
   <!-- A Trait is an experimentally measured phenotype for a QTL, which has ontology annotations -->
@@ -73,32 +46,34 @@
     <collection name="gwasResults" referenced-type="GWASResult" reverse-reference="trait"/>
   </class>
 
-   <!-- Annotatable for Publication support -->
+  <!-- GWAS is a study from a publication containing GWASResults, one per marker association -->
   <class name="GWAS" extends="Annotatable" is-interface="true">
     <attribute name="genotypes" type="java.lang.String"/>
     <attribute name="synopsis" type="java.lang.String"/>
     <attribute name="description" type="java.lang.String"/>
     <attribute name="genotypingPlatform" type="java.lang.String"/>
     <attribute name="genotypingMethod" type="java.lang.String"/>
     <reference name="organism" referenced-type="Organism"/>
+    <reference name="dataSet" referenced-type="DataSet"/>
     <collection name="results" referenced-type="GWASResult" reverse-reference="gwas"/>
-    <collection name="dataSets" referenced-type="DataSet"/>
   </class>
 
   <!-- GWASResult relates a marker to a measured trait with a p-value -->
-  <!-- Since a marker may have several genomic mappings, we store markerName here and populate markers with a post-processor -->
+  <!-- Since a marker may have several genomic mappings, we store markerName here -->
+  <!-- GWASResult.markers is populated with a post-processor -->
   <class name="GWASResult" is-interface="true">
     <attribute name="markerName" type="java.lang.String"/>
     <attribute name="pValue" type="java.lang.Double"/>
     <reference name="gwas" referenced-type="GWAS" reverse-reference="results"/>
     <reference name="trait" referenced-type="Trait" reverse-reference="gwasResults"/>
     <collection name="markers" referenced-type="GeneticMarker" reverse-reference="gwasResults"/>
-    <collection name="dataSets" referenced-type="DataSet"/>
+    <reference name="dataSet" referenced-type="DataSet"/>
   </class>
 
-  <!-- Genes can be associated with QTLs that span them, via overlapping markers, found in a post-processor -->
-  <class name="Gene" is-interface="true">
-    <collection name="qtls" referenced-type="QTL" reverse-reference="genes"/>
+  <!-- GeneticMarker collections populated by post-processors -->
+  <class name="GeneticMarker" is-interface="true">
+    <collection name="qtls" referenced-type="QTL" reverse-reference="markers"/>
+    <collection name="gwasResults" referenced-type="GWASResult" reverse-reference="markers"/>
   </class>
 
 </classes>
diff --git a/lis-genetic/src/main/resources/lis-genetic_keys.properties b/lis-genetic/src/main/resources/lis-genetic_keys.properties
@@ -15,9 +15,7 @@ Publication.key_doi=doi
 
 ## Annotatable
 Trait.key_primaryidentifier_dataset=primaryIdentifier,dataSet
-GeneticMap.key_primaryidentifier=primaryIdentifier
+LinkageGroup.key_primaryidentifier=primaryIdentifier
+QTLStudy.key_primaryidentifier=primaryIdentifier
+QTL.key_primaryidentifier_qtlstudy=primaryIdentifier,qtlStudy
 GWAS.key_primaryidentifier=primaryIdentifier
-
-## Other
-LinkageGroup.key_identifier_geneticmap=identifier,geneticMap
-QTL.key_identifier_linkagegroup=identifier,linkageGroup
diff --git a/lis-map/build.gradle b/lis-map/build.gradle
@@ -0,0 +1,28 @@
+sourceSets {
+    main {
+        java {
+            srcDirs = ['src/main/java', 'build/gen']
+        }
+        resources {
+            srcDirs = ['src/main/resources']
+        }
+    }
+    test {
+        java {
+            srcDirs = ['src/test/java']
+        }
+        resources {
+            srcDirs = ['src/test/resources']
+        }
+    }
+}
+
+dependencies {
+    compile group: 'org.intermine', name: 'bio-model', version: bioVersion //to read genomic_keyDefs.properties
+    bioModel group: 'org.intermine', name: 'bio-model', version: bioVersion, transitive: false
+    compile fileTree(dir: 'libs', include: '*.jar')
+}
+
+processResources {
+    from('.') { include ("*.properties")}
+}
diff --git a/lis-map/lis-map.properties b/lis-map/lis-map.properties
@@ -0,0 +1,6 @@
+compile.dependencies = intermine/objectstore/main,bio/core/main,\
+                       intermine/integrate/main
+
+have.file.custom.tgt = true
+
+converter.class = org.intermine.bio.dataconversion.MapFileConverter