WormBase-Caltech Weekly Calls January 2017
From WormBaseWiki
Jump to navigationJump to searchJanuary 5, 2017
BioCurator meeting
- Who from WormBase will go?
- Raymond (SObA), Chris (Community curation), Mary Ann, Wen would like to go
- Daniela maybe for micropublication
International C. elegans Meeting
- June 2017
- Whether all WB staff go depends on whether we have an all hands meeting before/after
- Plenty of things to present; AGR stuff etc.
Citace upload
- Prepare files by 10am on Tuesday, Jan 10
LEGO curation
- Kimberly working on different C. elegans models
- Working on several pathways simultaneously
- Should be in production soon (this year), fully evidenced
- Full benefit may not be apparent until AGR has LEGO fully integrated
- Working out how inferred annotations propagate from Noctua/LEGO model to GAF
- MGI creating many LEGO models
- How are curators, that went to training, using LEGO?
- Curation level view may not be amenable to general biologist/user viewing
- What is the best existing pathway display? WikiPathways? Protege?
- GO people (Seth, Chris M., Suzi etc.) are working on display, but have limited resources for this
- Maybe we could work with WikiPathways team?
- Pathway studio at RGD? Manually generated
- Should WB have a LEGO call? Yes, probably
- Curators may need access to Protein2GO for full curation/inference by sequence similarity
- How do we/should we resolve parallel pipelines for curating cellular component data (Go versus Expression pattern)?
- How do we accommodate specific versus general processes? E.g. specific versus general cell division
- We would like a mechanism for a LEGO curator to pull relevant data from GAF
- We would like curators to focus on pathway perspective of curation, where authors can submit experimental details with a potential automated or semi-automated mechanism to bridge the gap
Expression widget
- Daniela and Sibyl working on a more compact display of the expression data widget on gene pages
- Should we keep subcellular localization in the expression widget? Yes
SObA for Gene Ontology
- Raymond and Juancarlos working on
- First for WormBase, eventually want for AGR
- For GO data, will have three different roots (for each branch of GO); One DAG for each
- Could have phenotype graphs from particular perspective, e.g. via anatomy, via life stage, via GO, etc.
- Would be great to have a comparative view mechanism
January 12, 2017
Caltech LEGO Call
- Confirm time: Mondays, 11am PST (2pm EST/7pm GMT)
- Start January 23rd
- Format?
- Start with curation of one paper together?
- Curators then select individual topics?
Switching website to run off Datomic
- Juancarlos joining biweekly call on Wednesdays with Adam, Matt, and Sibyl
- Juancarlos working on Person page widgets currently
- Intellectual lineage graph will need to be rewritten in Clojure
- Web team is working out the file/directory structure to organize widget generation
- Is the Datomic database denormalized? To what extent? No, not really
- This affects how we might eventually write into Datomic from our curation database
- May be best for Caltech to keep a Postgres curation database in parallel to Datomic (and eventually sync nightly)
- This may also be best with AGR as well
Biocuration meeting 2017
- Who's going? Ranjana, Chris, Raymond, Mary Ann, Daniela (if she gets a talk)
- Ranjana - automated descriptions
- Raymond - SObA graph
- Chris - Genetic interactions
- Mary Ann - Community curation
- Daniela - Micropublications
Community curation presentation on AGR curator call
- Chris and Val Wood presented; minutes are here- AGR minutes
- WB community curation numbers
- 321 papers solicited and community curated (Phenotype OA)
- 417 papers community curated in total (Phenotype OA)
- 313 unique community curators (Phen OA); average 8 annotations per curator
- 64 papers solicited and community curated (RNAi OA)
- 74 papers community curated in total (RNAi OA)
- 61 unique community curators (RNAi OA); average 16 annotations per curator
- 3324 annotations total; 2969 valid and confirmed
- 2111 solicitations gone out; 16% response rate
Val presented some information on how complete contributors were:
- 1990s were getting 10 annotations/paper, in most recent years are now getting 50 annots/paper (years are publication years)
- Curation quality - users often omit a data type, but for what they do curate, they are generally very accurate, and often request more specific terms
Chris suspects ~20% authors are complete in their phenotype curation of their paper.
Pulling WB data into LEGO models
- Phenotype data requires a mapping step to map alleles and RNAi to genes
- How might we crowd source or get low cost curators to curate bulk data? Followed by expert curation for LEGO models
- May be best for curators to take on specific topics of biology
January 19, 2017
SObA graph for Gene Ontology
- Juancarlos and Raymond asking for feedback on GO SObA graph
- Kimberly: Can we create graphs for subsets of annotations (manual vs. automated)?
- Raymond: It's doable as a standalone graph, but may be difficult to make an option embedded within the core display
- May be a way to display IEA annotations separately from non-IEA annotations
- Raymond: Can develop a personal view for Kimberly to visualize manual vs. automated annotations
- Three root terms for GO, displayed side-by-side in the same graph; root nodes are different sizes and not necessarily displayed aligned
- May be able to force the root nodes to be the same size
- May want to show/hide individual roots; doable although unclear how difficult to implement
- GO function and component are likely to be smaller networks than process
- Pulling data from GAF into Solr
- Can this be adopted for the AGR portal? Yes
WormBase workshop at 2017 IWM
- Ranjana organizing
- Discuss ideas with Ranjana
- Have 1.5 - 2 hours
- Will want to discuss micropublications and community annotations
- We haven't had mirrors in a long time: discuss Datomic for people who want to query? Yes, brief announcement
User Survey
- We may want user surveys to prepare for grant writing, progress report
- Could send out to users by May?
- What to focus on: performance, data types, usability
- Could have multiple short surveys
- What one question would we want answers to help with design decisions?
- Do we want (eventually for AGR) a professional surveyor?
- Are our services meeting user needs?
- We can take a closer look at user logs; see what data, widgets, tools used most frequently (least frequently)
- We could create focus groups to get feedback from users on specific topics (30 minute sessions during poster sessions, for example); maybe at booth? during meal time?
- Would want to post a schedule with dedicated locations
- Could have WormBase hats, brightly colored ;), to indicate to people that we're WormBase staff
Disease data display on AGR portal gene page
- Working out a way to drill down from a summary view down to detailed table view
- Need to make sure it is intuitive and non-jargony for end users
- Yvonne Bradford drew up some mockups that represent a ribbon or heatmap display of annotations to disease categories for experimental models, etc.
- We do want a SObA graph; could possibly create a SObA graph that represents annotations to a gene AND all of its orthologs
First LEGO call
- This Monday (Jan 23) 11am Pacific
January 26, 2017
GO-CC annotations to paper term info in OA
- Daniela working on with Juancarlos
Disease working group
- Working to finalize the DAF (disease association file)
- Having extra meeting tomorrow
- Will then hand over DAF format to Kevin Howe (WB Data Quartermaster)
- In general, Kevin would prefer AGR files to be generated from the WB build process
Expression data working group?
- Daniela asking about an expression working group
- We don't have one currently, but one could be generated, although it may not make it into AGR portal 0.3, 0.6 or 1.0
Coherence, Consistency, and Continuity working group
- Working on a standard for table display in AGR portal
- Asking for standard functionality
- Table filtering, by column
- Table sorting, including sort order
- Paginated, choosing number of results per page
- Download options, tab-delimited and other formats
- Link to Intermine, to view, save and manipulate table and lists
- Currently WormBase has insufficient means to download some lists of results, for example downloading list of RNAi's that cause a lethal phenotype on the "lethal" phenotype page
- WB points users to FTP site, without specifying a file or how to parse
WormMine vs AGR Intermine
- Is Paulo working on shared/AGR Intermine or just WormMine? We think just WormMine
- There was an issue with making orthology calls which may have interfered with integrating worm data into common Intermine