Difference between revisions of "WormBase-Caltech Weekly Calls"

From WormBaseWiki
Jump to navigationJump to search
m
Line 69: Line 69:
 
*Who's going to handle the data? curate?
 
*Who's going to handle the data? curate?
 
*Michael? OK
 
*Michael? OK
 +
 +
 +
 +
==April 14, 2011==
 +
 +
-Gene Class Descriptions-
 +
*Concerns about maintenance and redundancy
 +
*Uma here for ~ 3 months
 +
*How many gene classes have alleles?
 +
*How many are named by phenotype rather than just molecular data?
 +
*How is this different from gene concise descriptions?
 +
*Should it be a summary of all gene concise descriptions of the class?
 +
*Things currently focused on:
 +
**using WormMart to look at genes in a class
 +
**pulls out all concise descriptions
 +
**look at similarities
 +
**interesting things to highlight
 +
*Gene concise descriptions vs class descriptions
 +
**Gene-centric vs Class-centric
 +
**Consolidating/pooling all concise descriptions from individual genes?
 +
*Going for maintenance-free statements
 +
*Potentially building an interface
 +
*Richard Durbin: development vs behavior?
 +
*Prioritization?
 +
*Focus on phenotype-based classes like UNC?
 +
*Factors for prioritization:
 +
**Numbers of genes curated
 +
**molecular vs phenotype-based
 +
**Amount of info currently available?
 +
**Historical points
 +
**Most actively worked currently? (most mentioned in last year's publications?)
 +
*Uma and Karen could communicate with Kimberly and Ranjana about
 +
*What is most efficient for Uma to focus on?
 +
*Uma can look at gene class description makes sense
 +
*Skip gene classes for which only one gene exists
 +
*GO term stats on each class?
 +
 +
 +
-Papers missing from Textpresso-
 +
*Issue: Genetics papers for GSA markup are missing from SVM analysis
 +
*Juancarlos' file on caprica
 +
*Discrepancy between papers on Textpresso and those gone through SVM
 +
*SVM doesn't pick up GSA papers
 +
*Generate a filtering to detect which ones have been missed by SVM
 +
*Michael looking into reasons why the pipeline isn't working
 +
*Tazendra vs Textpresso discrepancies?
 +
*Ruihua will process 56 missing papers retroactively
 +
*Still working on how to avoid this in the future

Revision as of 21:16, 14 April 2011

2009 Meetings


2011 Meetings

February

March


April 7, 2011

Transgene Model

  • On Wiki
  • Sent out to people
  • Have a look; report any concerns
  • Can follow on BitBucket; search for transgene; link to Wiki
  • No objections at Caltech; Karen will send to Paul Davis
  • Changes to ACE dumping script; Karen will talk to Juancarlos
  • Changes needed in OA (softer deadline than dump)


Interactions

  • Murky genetic interaction curation?
  • Err on the side of generality/trusting author statements
  • When in doubt, curate as "genetic interaction"
  • Chris is working on decision tree/pipeline for curation
  • Kimberly working on Physical Interaction model


BioGRID meeting at Princeton in May

  • Call in
  • What will Rose propose?


Expression Pattern Curation (Daniela/Wen)

  • Daniela sent out picture page for review
  • Expr Pattern OA wiki is in place:
  • As soon as Juancarlos is done with the modularization will start working on the code.
  • In the meanwhile Daniela will curate expression pattern writing .ace files
  • Expr_pattern OA should be ready by the next upload (May26th). (I really doubt this, parsing in data, writing dumpers, and checking it take a long time. Picture and Interaction each probably took longer than 2 months, and we're not starting Expr until May at the earliest -- Juancarlos)


Patch file/Interbuild (Raymond)

  • Developed good patch file
  • Tested patch file to update WS224 to WS225 - seems OK
  • Less than 5 minutes for upload
  • Testing now should be done by Todd/OICR team


Uma started

  • Working on concise descriptions of gene classes
  • Karen has reviewed with Uma; Uma is reading papers
  • Discussing details of descriptions
  • Inconsistencies/discrepancies of gene class names
  • >2400 gene classes
  • Can work on generating formula for this curation
  • Arun can help with automation
  • May need to get Uma an interface to enter data into postgres
  • Adapt concise description CGI for her? (probably write a whole new interface depending on goal -- Juancarlos)
  • Gene class name and a text field
  • Using Textpresso/WormMart output; sentence saver?


eggNOG data into citace?

  • Who's going to handle the data? curate?
  • Michael? OK


April 14, 2011

-Gene Class Descriptions-

  • Concerns about maintenance and redundancy
  • Uma here for ~ 3 months
  • How many gene classes have alleles?
  • How many are named by phenotype rather than just molecular data?
  • How is this different from gene concise descriptions?
  • Should it be a summary of all gene concise descriptions of the class?
  • Things currently focused on:
    • using WormMart to look at genes in a class
    • pulls out all concise descriptions
    • look at similarities
    • interesting things to highlight
  • Gene concise descriptions vs class descriptions
    • Gene-centric vs Class-centric
    • Consolidating/pooling all concise descriptions from individual genes?
  • Going for maintenance-free statements
  • Potentially building an interface
  • Richard Durbin: development vs behavior?
  • Prioritization?
  • Focus on phenotype-based classes like UNC?
  • Factors for prioritization:
    • Numbers of genes curated
    • molecular vs phenotype-based
    • Amount of info currently available?
    • Historical points
    • Most actively worked currently? (most mentioned in last year's publications?)
  • Uma and Karen could communicate with Kimberly and Ranjana about
  • What is most efficient for Uma to focus on?
  • Uma can look at gene class description makes sense
  • Skip gene classes for which only one gene exists
  • GO term stats on each class?


-Papers missing from Textpresso-

  • Issue: Genetics papers for GSA markup are missing from SVM analysis
  • Juancarlos' file on caprica
  • Discrepancy between papers on Textpresso and those gone through SVM
  • SVM doesn't pick up GSA papers
  • Generate a filtering to detect which ones have been missed by SVM
  • Michael looking into reasons why the pipeline isn't working
  • Tazendra vs Textpresso discrepancies?
  • Ruihua will process 56 missing papers retroactively
  • Still working on how to avoid this in the future