Archived assemblies: Difference between revisions

From genomewiki
Jump to navigationJump to search
Line 16: Line 16:


==dbDbArch==
==dbDbArch==
If a line in dbDb is no longer needed for another assembly's net, chain, or conservation track, it can be removed and added to hgcentral.dbDbArch instead.  The dbDbArch table is present to support hgLiftOver and hgConvert, which will continue to work for archived assemblies as long as the pertinent information in /gbdb is still available.
All archived assemblies should be represented by a row in hgcentral.dbDbArch.  These rows can be copied from dbDb.  If a row in dbDb is no longer needed for another assembly's net, chain, or conservation track, it can be removed from dbDb.  The dbDbArch table is present to support hgLiftOver and hgConvert, which will continue to work for archived assemblies as long as the pertinent files in /gbdb/<db> are still available.


=Mirrors=
=Mirrors=

Revision as of 21:26, 16 September 2008

Older assemblies are archived on the archives server, at genome-archive.cse.ucsc.edu. However, some metadata about the archived assemblies is kept in the hgcentral database in order to offer liftOver to and from the older assemblies.

NOTE: the following setup describes the database setup that should take effect with the v172 cgis. The old setup used a separate database, hgcentarchive, with its own dbDb table. This entire database is being replaced by the single table dbDbArch in hgcentral.

hgcentral setup

liftOverChain

Entries in liftOverChain should be kept for archived assemblies in order to keep nets/chains from other browsers that point to the archived assemblies from breaking. Also, the entries make it possible for hgLiftOver and hgConvert to continue to work for archived assemblies on the main site (not genome-archive).

dbDb

A line in dbDb for an archived assembly should be maintained if:

1. Any other organism's nets and chains are still pointing to it.

2. The conservation track of another organism was built using the archived assembly. The conservation track needs the dbDb entry in order to find the organism name for the hgc page (it uses it to write "Rhesus" rather than "rheMac1", for instance).

dbDbArch

All archived assemblies should be represented by a row in hgcentral.dbDbArch. These rows can be copied from dbDb. If a row in dbDb is no longer needed for another assembly's net, chain, or conservation track, it can be removed from dbDb. The dbDbArch table is present to support hgLiftOver and hgConvert, which will continue to work for archived assemblies as long as the pertinent files in /gbdb/<db> are still available.

Mirrors

The old system of using a separate database for archived assemblies consistently caused problems for mirror sites on hgLiftOver and hgConvert. These cgis gave an error message like:

Please set archivecentral options in the hg.conf file.

This problem should be eliminated with the new dbDbArch table in hgcentral.

Mirror sites should expect hgLiftOver and hgConvert to work for any assemblies for which they have sequence data available in /gbdb and an entry in the hgcentral.liftOverChain table. Partial mirrors should only see the assemblies they have chosen to mirror in the drop-down menus of hgLiftOver and hgConvert.

Mirrors can safely remove these lines from their hg.conf files, if present:

 archivecentral.host=
 archivecentral.db=
 archivecentral.user=
 archivecentral.password=
 
 archivebackup.host=
 archivebackup.db=
 archivebackup.user=
 archivebackup.password=

(providing the cgis are v172 or later).

Archive dated Aug 25, 2008

Details [1] : of seven assemblies archived 2008-08-25.