Updated for ChEMBL 09

egonw · Mar 17, 2011 · 3fa2dfd · 3fa2dfd
1 parent e79dbd1
commit 3fa2dfd
Show file tree

Hide file tree

Showing 6 changed files with 188 additions and 49 deletions.
diff --git a/.gitignore b/.gitignore
@@ -1,2 +1,2 @@
 vars.php
-
+*~
diff --git a/README.markdown b/README.markdown
@@ -11,30 +11,30 @@ The data is copyrighted by them, and licensed CC-BY-SA:
 as explained on:
 
   http://www.ebi.ac.uk/chembldb/
-  ftp://ftp.ebi.ac.uk/pub/databases/chembl/releases/chembl_06/LICENSE
+  ftp://ftp.ebi.ac.uk/pub/databases/chembl/releases/chembl_09/LICENSE
 
 These scripts were tested against version 06 of ChEMBL, as downloaded from:
 
-  ftp://ftp.ebi.ac.uk/pub/databases/chembl/releases/chembl_06/
+  ftp://ftp.ebi.ac.uk/pub/databases/chembl/releases/chembl_00/
 
 # Requirements
 
-ChEMBL 06.
+ChEMBL 09.
 
 # Installation
 
 The scripts expect a script only readble by the server software called vars.php, with content like:
 
-<?php
+    <?php
 
-$db = 'chembl_06';
-$user = 'user';
-$pwd = 'secret';
+    $db = 'chembl_09';
+    $user = 'user';
+    $pwd = 'secret';
 
-// use the next line to limit the output 
-// $limit = ' LIMIT 5';
-$limit = '';
+    // use the next line to limit the output 
+    // $limit = ' LIMIT 5';
+    $limit = '';
 
-?>
+    ?>
 
 to access the MySQL database with the ChEMBL content.
diff --git a/activities.php b/activities.php
@@ -33,17 +33,19 @@
   echo "act:a" . $row['activity_id'] . " a :Activity ;\n";
   echo " :extractedFrom res:r" . $row['doc_id'] . " ;\n";
   echo " :onAssay ass:a" . $row['assay_id'] . " ;\n";
-  $chebi = mysql_query("SELECT DISTINCT * FROM compounds WHERE molregno = \"" . $row['molregno'] . "\"");
+  $chebi = mysql_query("SELECT DISTINCT * FROM molecule_dictionary WHERE molregno = \"" . $row['molregno'] . "\"");
   if ($chebiRow = mysql_fetch_assoc($chebi)) {
     echo " :forMolecule mol:m" . $chebiRow['chebi_id'] . " ;\n";
   }
   if ($row['relation']) {
     if ($row['relation'])
       echo " :relation \"" . $row['relation'] . "\" ;\n";
   }
-  echo " :standardValue \"" . $row['standard_value'] . "\" ;\n";
-  echo " :standardUnits \"" . $row['standard_units'] . "\" ;\n";
-  echo " :type \"" . $row['standard_type'] . "\" .\n";  
+  if ($row['standard_value']) {
+    echo " :standardValue \"" . $row['standard_value'] . "\"^^xsd:float ;\n";
+    echo " :standardUnits \"" . $row['standard_units'] . "\" ;\n";
+    echo " :type \"" . $row['standard_type'] . "\" .\n";
+  }
   flush();
 }
 

diff --git a/chembl.owl b/chembl.owl
@@ -0,0 +1,112 @@
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+@prefix owl: <http://www.w3.org/2002/07/owl#> .
+
+@prefix bibo: <http://purl.org/ontology/bibo/> .
+
+@prefix : <http://pele.farmbio.uu.se/chembl/onto/#> .
+
+:Drug owl:equivalentClass :Compound ;
+  a owl:Class .
+
+:SmallMolecule rdfs:subClassOf :Drug .
+:Protein rdfs:subClassOf :Drug .
+:Cell rdfs:subClassOf :Drug .
+:Oligosaccharide rdfs:subClassOf :Drug .
+:Oligonucleotide rdfs:subClassOf :Drug .
+:Antibody rdfs:subClassOf :Drug .
+
+:Activity a owl:Class .
+:Assay a owl:Class .
+:Target a owl:Class .
+
+:extractedFrom a owl:ObjectProperty ;
+  rdfs:domain :Activity ;
+  rdfs:range bibo:Document .
+
+:onAssay a owl:ObjectProperty ;
+  rdfs:domain :Activity ;
+  rdfs:range :Assay .
+
+:forMolecule a owl:ObjectPropery ;
+  rdfs:domain :Activity ;
+  rdfs:range :Drug .
+
+:relation a owl:DatatypeProperty ;
+  rdfs:domain :Activity .
+
+:standardValue a owl:DatatypeProperty ;
+  rdfs:domain :Activity .
+
+:standardUnits a owl:DatatypeProperty ;
+  rdfs:domain :Activity .
+
+:type a owl:DatatypeProperty ;
+  rdfs:domain :Activity .
+
+:hasDescription a owl:DatatypeProperty ;
+  rdfs:domain :Assay .
+
+:extractedFrom a owl:ObjectProperty ;
+  rdfs:domain :Assay ;
+  rdfs:range bibo:Document .
+
+:organism a owl:DatatypeProperty ;
+  rdfs:domain :Assay .
+
+:hasTarget a owl:ObjectProperty ;
+  rdfs:domain :Assay ;
+  rdfs:range  :Target .
+
+:hasConfScore a owl:DatatypeProperty ;
+  rdfs:domain :Assay .
+
+:hasAssayType a owl:ObjectProperty ;
+  rdfs:domain :Assay .
+
+:hasTargetType a owl:ObjectProperty ;
+  rdfs:domain :Target .
+
+:organism a owl:DatatypeProperty ;
+  rdfs:domain :Target .
+
+:hasDescription a owl:DatatypeProperty ;
+  rdfs:domain :Target .
+
+:hasKeyword a owl:DatatypeProperty ;
+  rdfs:domain :Target .
+
+:sequence a owl:DatatypeProperty ;
+  rdfs:domain :Target .
+
+:hasTaxonomy a owl:ObjectProperty ;
+  rdfs:domain :Target ;
+  rdfs:range  <http://bio2rdf.org/uniprot_resource:Taxon> .
+
+:classL1 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 1 classification".
+
+:classL2 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 2 classification".
+
+:classL3 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 3 classification".
+
+:classL4 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 4 classification".
+
+:classL5 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 5 classification".
+
+:classL6 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 6 classification".
+
+:classL7 a owl:DatatypeProperty ;
+  rdfs:domain :Target ;
+  rdfs:label "Level 7 classification".
diff --git a/compounds.php b/compounds.php
@@ -45,46 +45,71 @@
 #  echo "</rdf:Description>\n";
 #}
 
-$allIDs = mysql_query("SELECT * FROM compounds, compound_properties WHERE compounds.molregno = compound_properties.molregno" . $limit);
+$allIDs = mysql_query("SELECT * FROM compound_records, compound_properties WHERE compound_records.molregno = compound_properties.molregno" . $limit);
 
 $num = mysql_numrows($allIDs);
 
 while ($row = mysql_fetch_assoc($allIDs)) {
-  echo "mol:m" . $row['chebi_id'] . " a :Compound ;\n";
-  #if ($row['inchi']) {
-  if (false)
-    echo " = <http://rdf.openmolecules.net/?" . $row['inchi'] . "> ;\n";
-    echo " chem:inchi \"" . $row['inchi'] . "\" ;\n";
-  }
-  #if ($row['inchi_key'])
-  #  echo " chem:inchikey \"" . $row['inchi_key'] . "\" ;\n";
-  if ($row['canonical_smiles'])
-    echo " chem:smiles \"" . $row['canonical_smiles'] . "\" ;\n";
-
-  #foreach ($descs as $value => $type) {
-    #if ($row[$row[$value]]) {
+  $chebi = mysql_query("SELECT DISTINCT * FROM molecule_dictionary WHERE molregno = \"" . $row['molregno'] . "\"");
+  if ($chebiRow = mysql_fetch_assoc($chebi)) {
+    echo "mol:m" . $chebiRow['chebi_id'] . " a ";
+    if ($chebiRow['molecule_type']) {
+      if ($chebiRow['molecule_type'] = "Small molecule") {
+        echo ":SmallMolecule ;\n";
+      } else if ($chebiRow['molecule_type'] = "Protein") {
+        echo ":Protein ;\n";
+      } else if ($chebiRow['molecule_type'] = "Cell") {
+        echo ":Cell ;\n";
+      } else if ($chebiRow['molecule_type'] = "Oligosaccharide") {
+        echo ":Oligosaccharide ;\n";
+      } else if ($chebiRow['molecule_type'] = "Oligonucleotide") {
+        echo ":Oligonucleotide ;\n";
+      } else if ($chebiRow['molecule_type'] = "Antibody") {
+        echo ":Antibody ;\n";
+      } else {
+        echo ":Drug ;\n";
+      }
+    } else {
+      echo ":Drug ;\n";
+    }
+    #if ($row['inchi']) {
     if (false) {
-      echo "  <bodo:hasDescriptorValue>\n";
-      echo "    <bodo:DescriptorValue>\n";
-      echo "      <bodo:hasPart>\n";
-      echo "        <bodo:DescriptorValuePoint>\n";
-      echo "          <bodo:hasValue rdf:datatype=\"http://www.w3.org/2001/XMLSchema#" .
-           $type . "\">" . $row[$value] . "</bodo:hasValue>\n";
-      echo "          <bodo:valuePointFor rdf:resource=\"" . $ns . "descriptorId=" . md5($value) . "\" />\n";
-      echo "        </bodo:DescriptorValuePoint>\n";
-      echo "      </bodo:hasPart>\n";
-      echo "    </bodo:DescriptorValue>\n";
-      echo "  </bodo:hasDescriptorValue>\n";
+      echo " = <http://rdf.openmolecules.net/?" . $row['inchi'] . "> ;\n";
+      echo " chem:inchi \"" . $row['inchi'] . "\" ;\n";
+    }
+    #if ($row['inchi_key'])
+    #  echo " chem:inchikey \"" . $row['inchi_key'] . "\" ;\n";
+    $structs = mysql_query("SELECT DISTINCT * FROM compound_structures WHERE molregno = " . $row['molregno']);
+    while ($struct = mysql_fetch_assoc($structs)) {
+      if ($struct['canonical_smiles'])
+        echo " chem:smiles \"" . $struct['canonical_smiles'] . "\" ;\n";
     }
-  #}
 
-  $names = mysql_query("SELECT DISTINCT * FROM compound_synonyms WHERE molregno = " . $row['molregno']);
-  while ($name = mysql_fetch_assoc($names)) {
-    if ($name['synonyms'])
-      echo " dc:title \"" . str_replace("\"", "\\\"", $name['synonyms']) . "\" ;\n";
-  }
+    #foreach ($descs as $value => $type) {
+      #if ($row[$row[$value]]) {
+      if (false) {
+        echo "  <bodo:hasDescriptorValue>\n";
+        echo "    <bodo:DescriptorValue>\n";
+        echo "      <bodo:hasPart>\n";
+        echo "        <bodo:DescriptorValuePoint>\n";
+        echo "          <bodo:hasValue rdf:datatype=\"http://www.w3.org/2001/XMLSchema#" .
+            $type . "\">" . $row[$value] . "</bodo:hasValue>\n";
+        echo "          <bodo:valuePointFor rdf:resource=\"" . $ns . "descriptorId=" . md5($value) . "\" />\n";
+        echo "        </bodo:DescriptorValuePoint>\n";
+        echo "      </bodo:hasPart>\n";
+        echo "    </bodo:DescriptorValue>\n";
+        echo "  </bodo:hasDescriptorValue>\n";
+      }
+    #}
+
+    $names = mysql_query("SELECT DISTINCT * FROM molecule_synonyms WHERE molregno = " . $row['molregno']);
+    while ($name = mysql_fetch_assoc($names)) {
+      if ($name['synonyms'])
+        echo " dc:title \"" . str_replace("\"", "\\\"", $name['synonyms']) . "\" ;\n";
+    }
 
-  echo " = <http://bio2rdf.org/chebi:" . $row['chebi_id'] . "> .\n";
+    echo " = <http://bio2rdf.org/chebi:" . $chebiRow['chebi_id'] . "> .\n";
+  }
 }
 
 ?>
diff --git a/docs.php b/docs.php
@@ -16,7 +16,7 @@
 @prefix res: <http://rdf.farmbio.uu.se/chembl/resource/> .
 @prefix mol: <http://rdf.farmbio.uu.se/chembl/molecule/> .
 @prefix ass: <http://rdf.farmbio.uu.se/chembl/assay/> .
-@prefix jrn: <http://rdf.farmbio.uu.se/chembl/assay/> .
+@prefix jrn: <http://rdf.farmbio.uu.se/chembl/journal/> .
 
 <?php