PhenCode: Example of adding a new data source
In order to load the data into the Locus Variants track, it must be
rearranged into tab-delimited files that match the track's table
schema, with one file for each table.
Other tables used for producing displays
The main base table is called gv ("genome variation"). It
contains one entry (row) for each mutation. Each mutation needs a
unique identifier, which is created by concatenating the abbreviation
for the database with the database's unique identifier for the mutation.
If the database does not have unique identifiers for its mutations, then
the HGVS-style names are used.
The gvPos[Build] table stores the position fields that place each
mutation onto the track. If the data is to appear on the Genome Browser
for several assemblies (genome builds), then there will be a version of
this table for each one, since the coordinates will likely be different.
The gvLink table contains values for building links back to the
data source and to other databases if available. The template
describing how to build the links is stored in a flat file
links.ra) in the UCSC source code tree.
The gvSrc table contains the full name of the data source, and
also distinguishes between LSDBs and genome-wide data sets.
The gvAttr table stores all the other data
that is displayed on the details page for each mutation. This also
opens up opportunities for querying across multiple data sources
simultaneously, by using common attributes stored here.
attrType field is used to group
similar types of data and to control how the attribute is displayed on
the details page.
- The linkBuilder table stores the information on the base
URL and information needed to generate links back to the LSDBs or other
- The srcDetails table stores the details about the source such
as the logo and names for credits.
- The gvAttributes table stores the categories, display order, and
other information on how to display the attributes.
The precise schema definitions for these tables can be viewed
return to FAQ