Ensembl QA

From Genecats
Revision as of 23:09, 24 April 2013 by Brianlee (talk | contribs)
Jump to navigationJump to search

There are periodic updates to all Ensembl Genes tracks in all dbs. Here is a list of things to do in addition to the regular QA you would do normally for a track:

  • Run qaEnsGenes.csh
    • Run it before you push to beta, as it compares dev to beta. You'll lose that comparison if you push too soon.
  • All Ensembl builds are documented in the single file:
   src/hg/makeDb/doc/makeEnsembl.txt
  • If it is the first assembly, however, it will rather be documented in the assembly.txt make document.
   /kent/src/hg/makeDb/doc/nameAssembly.txt
  • You can find all ensembl gene builds in:
   /hive/data/genomes/<db>/bed/ensGene.NN/
where NN is the ensembl version number.
  • To get the links to go to the correct place at Ensembl, you will have to check for the appropriate rows in trackVersion table in hgFixed database
    • The latest entry in the hgFixed.trackVersion table should have "current" in the dateReference field and the previous entry needs to have the dateReference field updated to whatever month the last Ensembl release was dated.
    • Before you push the hgFixed.trackVersion table, check (using the compareWholeTable.csh script) that you aren't pushing changes that aren't ready to go out yet. For instance, if there are changes to an unrelated track, such as cosmic, you will need to check with the person who made the changes to see if it is okay to push the changes. If it is not, the timing of the push should be coordinated or the trackVersion table should be edited on mysqlbeta to exclude the changes that are not ready to go out.
  • New: You also need to update the blastTabs if the zebrafish assembly used in hgGene got new Ensembl Genes. Talk to Brooke or Mary if this is case.