bde6c9904a5b213edd75c32f811b8879bed09e1f
cline
  Sun Jan 22 18:13:02 2012 -0800
Addressing Redmine bug # 6661 by changing the wording for the chromStart and chromEnd fields to indicate that they correspond to txStart and txEnd respectively for + strand genes, not for all genes
diff --git src/hg/lib/knownCanonical.sql src/hg/lib/knownCanonical.sql
index 12efa78..a2ba824 100644
--- src/hg/lib/knownCanonical.sql
+++ src/hg/lib/knownCanonical.sql
@@ -1,18 +1,16 @@
 # knownCanonical.sql was originally generated by the autoSql program, which also 
 # generated knownCanonical.c and knownCanonical.h.  This creates the database representation of
 # an object which can be loaded and saved from RAM in a fairly 
 # automatic way.
 
 #Describes the canonical splice variant of a gene
 CREATE TABLE knownCanonical (
     chrom varchar(255) not null,	# Chromosome
-    chromStart int not null,	# Start position (0 based). Corresponds to txStart
-    chromEnd int not null,	# End position (non-inclusive). Corresponds to txEnd
+    chromStart int not null,	# Start position (0 based). Corresponds to txStart for + strand genes
+    chromEnd int not null,	# End position (non-inclusive). Corresponds to txEnd for + strand genes
     clusterId int not null,	# Which cluster of transcripts this belongs to in knownIsoforms
     transcript varchar(255) not null,	# Corresponds to knownGene name field.
-    protein varchar(255) not null,	# SwissProt ID of associated protein.
+    protein varchar(255) not null,	# SwissProt ID or more recently UCSC ID of associated protein.
               #Indices
-    UNIQUE(clusterId),
-    INDEX(transcript(12)),
-    INDEX(protein(12))
+    PRIMARY KEY(chrom)
 );