inprogress

diff --git a/forester/resources/phyloxml_schema/1.10/phyloxml.xsd b/forester/resources/phyloxml_schema/1.10/phyloxml.xsd
index 525bc69..5512426 100644 (file)
--- a/forester/resources/phyloxml_schema/1.10/phyloxml.xsd
+++ b/forester/resources/phyloxml_schema/1.10/phyloxml.xsd
@@ -16,7 +16,7 @@
 <!--                                                             -->\r
 <!-- www.phyloxml.org                                            -->\r
 <!-- Version: 1.16                                               -->\r
-<!-- Last modified: 2013.09.29 by Christian M Zmasek             -->\r
+<!-- Last modified: 2013.09.30 by Christian  Zmasek              -->\r
 <!--                                                             -->\r
 <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:phy="http://www.phyloxml.org"\r
    targetNamespace="http://www.phyloxml.org" elementFormDefault="qualified" attributeFormDefault="unqualified">\r
@@ -192,17 +192,18 @@
    <xs:complexType name="Sequence">\r
       <xs:annotation>\r
          <xs:documentation> Element Sequence is used to represent a molecular sequence (Protein, DNA, RNA) associated\r
-            with a node. 'symbol' is a short (maximal 30 characters) symbol of the sequence (e.g. 'ACTM') whereas\r
-            'name' is used for the full name (e.g. 'muscle Actin'). 'location' is used for the location of a sequence on\r
-            a genome/chromosome. The actual sequence can be stored with the 'mol_seq' element. Attribute 'type' is used\r
-            to indicate the type of sequence ('dna', 'rna', or 'protein'). One intended use for 'id_ref' is to link a\r
-            sequence to a taxonomy (via the taxonomy's 'id_source') in case of multiple sequences and taxonomies per\r
-            node. </xs:documentation>\r
+            with a node. 'symbol' is a short (maximal 20 characters) symbol of the sequence (e.g. 'ACTM') whereas\r
+            'name' is used for the full name (e.g. 'muscle Actin'). 'gene_name' can be used when protein and gene names differ.\r
+            'location' is used for the location of a sequence on a genome/chromosome. The actual sequence can be stored with the \r
+            'mol_seq' element. Attribute 'type' is used to indicate the type of sequence ('dna', 'rna', or 'protein'). \r
+            One intended use for 'id_ref' is to link a sequence to a taxonomy (via the taxonomy's 'id_source') in case \r
+            of multiple sequences and taxonomies per node. </xs:documentation>\r
       </xs:annotation>\r
       <xs:sequence>\r
          <xs:element name="symbol" type="phy:SequenceSymbol" minOccurs="0"/>\r
          <xs:element name="accession" type="phy:Accession" minOccurs="0"/>\r
          <xs:element name="name" type="xs:token" minOccurs="0"/>\r
+         <xs:element name="gene_name" type="xs:token" minOccurs="0"/>\r
          <xs:element name="location" type="xs:token" minOccurs="0"/>\r
          <xs:element name="mol_seq" type="phy:MolSeq" minOccurs="0"/>\r
          <xs:element name="uri" type="phy:Uri" minOccurs="0" maxOccurs="unbounded"/>\r
@@ -217,7 +218,7 @@
    </xs:complexType>\r
    <xs:simpleType name="SequenceSymbol">\r
       <xs:restriction base="xs:token">\r
-         <xs:pattern value="\S{1,30}"/>\r
+         <xs:pattern value="\S{1,20}"/>\r
       </xs:restriction>\r
    </xs:simpleType>\r
    <xs:complexType name="MolSeq">\r
@@ -249,6 +250,7 @@
       <xs:simpleContent>\r
          <xs:extension base="xs:token">\r
             <xs:attribute name="source" type="xs:token" use="required"/>\r
+            <xs:attribute name="comment" type="xs:token"/>\r
          </xs:extension>\r
       </xs:simpleContent>\r
    </xs:complexType>\r