Reference Genome April 2010-monthly (Archived)

From GO Wiki
Jump to: navigation, search

Monthly conference call

The monthly conference call has been canceled to give people time to get the working groups started for the GO annotation camp.

Testing import of PAINT-generated GAF files by MODs

Reference_genome-TEST GAF file

PAINT SOP meeting

We held a meeting in Princeton with Kara Dolinski, Mike Livestone, Mary Dolan, Li Ni, Pascale Gaudet, Paul Thomas, Ed Lee, as well as Suzi Lewis and Judy Blake (both attending remotely). A lot of the SOPs have now been defined. The minutes are here: RefG_Princeton_April_12-13_2010

PAINT beta 29

Has several fixes and improvements that were discussed during the Princeton meeting:

  • Literature linkouts for Wormbase and ZFIN have been fixed
  • Radio buttons showing aspects have been reordered to: Molecular Function / Cellular Component / Biological Process to reflect the order of the annotation process
  • Added scrolling to Evidence panel
  • NOTs due to rapid divergence are not output to the export GAF (but are maintained in the save GAF to allow reloading session)
  • Added configurable (currently through a XML file) list of terms to exclude from the Annotation Matrix. Currently these terms are: GO:0005515 (protein binding), GO:0005488, (binding), and GO:0019899 (enzyme binding). You can manually edit the XML file to add or remove terms to be excluded.
  • Fixed scrolling of ruler in MSA view
  • Other minor bug fixes

PAINT beta 28

  • GAFs generated are now GAF 2.0 compliant -- note that there was a bug in the previous versions where generated GAFs had the wrong number of columns, so if you have any GAFs that you have saved in previous versions of PAINT, they won't work with the current version (to fix the files is fairly trivial, so if you have any issues, feel free to contact me). The GAF parser supports both GAF 1.0 and GAF 2.0 formats.
  • Issues with timing out connections to the GO database have *hopefully* been resolved. The side effect of this issue was that PAINT instances that were idle for long periods of time would hang, causing the user to need to restart the application. Please provide feedback on whether PAINT is still hanging after long periods of time (although with how much faster it now is, you might not just let PAINT sit while you go grab a cup of coffee when waiting for PAINT to propagate annotations -- sorry Paul =P ).

PAINT beta 24-27

We've just released beta 27, but first is a summary of the changes between 24 and 26:

  • Importing GAF files and the annotating is MUCH faster now. We're talking orders of magnitude here, folks. Some pretty clever tricks were used to speed things up, but since a lot of the code had to be rewritten, hopefully it didn't break anything. Let us know if something has gone awry after the massive rewriting.
  • Hyperlinks for E.coli genes now work (they were broken after we had to change the database from ECOLI to EcoCyc as expected in the GO database)
  • Fixed issues where experimental evidence from the GO database sometimes wouldn't show up in PAINT. Changed the behavior so that all experimental evidence shows up, even if they're to the same term (as long as the SOURCE and/or ECO are different).
  • Fixed issues with weird colored nodes when restoring data from a GAF that contained NOT annotations in a particular aspect but not other(s)
  • Fixed scrolling issues with MSA viewer when using the horizontal scroll bar (worked fine when using the mouse wheel).

Phew, now on to beta 27...

  • Fixed some issues with how associations are generated that was preventing multiple annotations to certain nodes
  • After some feedback from the GO meeting and discussion with Paul, we've decided to change how NOTs are displayed in the annotation matrix. In the past, NOTs would display up the the GO graph as long as there was no positive annotation to contradict. That turned out to be confusing, as NOTs don't propagate up, but rather down. So the behavior has been changed to reflect that. Implied NOTs are only applied to descendants. Since this is a new feature, I'd like to get some feedback to make sure that it's behaving correctly. This also brings up another question. Should we expand the headers (GO terms) in the matrix to display all descendant terms to currently annotated terms? We probably wouldn't annotate to those terms directly, but they would show the NOT downwards propagation. The downside of that is it will fatten up the matrix even further.

Back to Reference_Genome_Annotation_Project#Progress_Reports