WormBase-Caltech Weekly Calls September 2013
From WormBaseWikiJump to navigationJump to search
- 1 September 5, 2013
- 2 September 12, 2013
- 3 September 19, 2013
September 5, 2013
- Wen contacted Matt Hibbs
- Matt will not maintain SPELL in near future
- Undergrad at Caltech to maintain/upgrade (Ruby 2 -> Ruby 3)?
- Connect Matt with undergrad (work study?)
- Cynthia(?) at Caltech may know Ruby (needed for interface)
- Communicate with Mike Cherry / SGD as they use SPELL as well
- Does SGD have a SPELL developer?
- Java experience needed to maintain backend
- Wen found that many authors don't publish lists of clustered genes
- We could take in the data
Gene ACE Dumps
- Kimberly took a look at what Michael sent us
- Looked at gene objects specifically
- More in the full Gene ACE that we could make use of than what was included in the 'nightly' dump from Michael
- Could other curators take a look at what could be used/requested from GeneACE?
- Karen and Juancarlos will discuss for variations today
- Examples: CDS to Pseudogene changes, gene-to-variation mappings/pairings
Discussions on Value of Manual Curation
- What would WormBase look like without any manually curated data?
- Can concise descriptions be automated in any way? "This gene encodes ... It is involved in [GO process]..." etc.
- How are the concise description forms working? In progress...
- Gold-standard enhancers from Margie's list of papers
- Xiaodong will upload for WS240
- Interaction model changed to update genetic interaction terms
- Also, "Electrophoretic Mobility Shift Assay" has been added as a 'Physical Interaction Detection Method'
- Interaction OA and dumper has been updated and is working properly
- Abby has setup display of historical genes in Overview widgets of relevant objects (Interaction, Expr_pattern, Antibody, Transgene, Process)
- Looks good on staging
Upgrading transgene curation
- Capturing related sequences or sequence features and displaying on GBrowse
- Single molecule FISH data: Use case
- Look in GBrowse to see probe sequence vs transgene sequence; find discrepancy in expression patterns
- Tim Schedl would like to see what transgenes have been used for in vitro (purified from worms) experiments (e.g. physical interactions)
- Would like to capture epitope tags like FLAG, MYC, HIS, etc.
- Capturing what does or does not rescue a mutation
- Association with site-of-action data?
In silico phylogenetic tree data
- No current appropriate home for it in WormBase
- Wen and Karen working on how to house the data
- Data should go on Process pages
Process curation and associations
- Can we automatically infer associations to a process by, for example, a gene's mutant phenotype?
- What automatic connections can be made to processes?
- Do we 'hardcode' phenotype-to-process connections?
- Interaction widget with Cytoscape now functional on Process pages (on staging)
September 12, 2013
- RNAi.org movies need referencing
- May require adding a Database tag to the ?Movie model
- We may be able to add the RNAi.org movie URL to the Movie object Remark field
- Chris will work it out with Todd
Crispr-CAS9 Genome Editing methods
- Will add to the Reagents page (will be under the Resources menu, not there on live site currently)
- URL is: http://www.wormbase.org/resources/reagents
- Curators can edit or create new widgets if logged in
- Ranjana will add two new widgets: "Third Party Tools", "Methods"
- Will add Ortholist to the Third Party Tools
- Will add Crispr-CAS9 links to Methods
GeneACE Nightly Dump
- What data do we want to get from Hinxton (in general) on a daily/nightly basis?
- We will wait for Kimberly to settle on what we want
- Raymond establishing a WormBase development environment that Juancarlos can use
- Troubleshooting installation of Perl libraries
- Also, working to Establish a development environment for the WormBase Ontology Browser server
- With our own server we can supply WormBase-specific ontologies
- When can we/should we request new data for WormMine?
- JD is working on getting new data in; there is a Wiki page for requesting new data; we can add new data there
Cytoscape plugin update
- New viewer, has a different feel
- Need to click and hold to pan and scroll to zoom
- Hopefully adding back visible tools soon
- Implementing on Process pages and for Cell lineage browsing
- Cell lineage browsing
- First tried whole (embryonic) lineage but was too slow because too many nodes
- Now have a subtree rendering with 4 levels of the tree (great-grandparents, grandparents, parents, and children)
- May like to have full lineage at high resolution to zoom into for browsing a static image
- Process pages
- Cytoscape view of Process pages has all interactions related to the process
- Phenotypes for genetic interactions now listed in the legend
ORCID IDs for WormBase Persons
- We should incorporate ORCID IDs
- Can we do a reciprocal exchange of persons? Opt-in exchange?
- Should we ask WormBase users for their ORCID IDs?
- Came up while discussing Reuter's indexing (of gene name authors)
- We will add a request for ORCID IDs on the WormBase form for creating or editing WBPerson info
September 19, 2013
Caltech Development Environment
- Issues with running a local mirror
- Less tenable option; build and staging environments are distributed; Nothing sitting on a single machine
- Data size: amount of storage required; amount of time required for the build
- Build process lengthy; need to distribute the task or wait a week
- Todd can give an AMI (?) ID , letting Raymond create an instance
- Raymond wouldn't need to build the GFF database
- EC2 cost? Depends on how
- Storage is biggest cost; can minimize (use our MySQL hosts), still need to run ACE
- Run local storage off of femoral(?)
- EC2 instance:
- Running 24 hours a day? Use 25%?
- Let's say 24 hour EC2 instance, large size, heavy utilization, upfront cost of 1 year (with purchased reserved instance): ~$1166
- Plus bandwidth and storage costs
- Total cost : ~$1500 per year
- Environments change name to update: Development --> Staging --> Production
- 3rd option: setting Raymond up on development service (?)
- May be easiest approach
- Todd opts for this approach
- Raymond and Juancarlos can have their own logins
- Todd is drafting the NAR manuscript
- Opting for the short format (2-3 pages); not many people read it; geared towards database people
- Todd wants to describe new WormBase approach in a different paper
- What are the big impact changes on the curatorial side?
- Curators could collect points and send them to Todd
- Deadline is the 1st week in October
- Curation section about 1 page; Focus on direction: processes, pathways, topic-based curation
- Curators can focus on building off of a 'standard' pathway diagram from a review article, for example
- Can start with a review, WormBook chapter
- What coverage are we shooting for? 90%?
- Start a WikiPathway; curators can build off of that
- Need to improve:
- How we choose a topic?
- How we find the papers?
- What we capture? How we converge our efforts?
- Karen made a process for each type of stress response
- We can start with Misfolded (Unfolded) Protein Response (related terms: "proteotoxicity", "ER stress")
- Curators can keep notes about topic curation
- We want to consider GO terms, existing WB Processes, existing WikiPathways
- Karen will be primarily responsible for Process and WikiPathway curation
GeneACE Nightly Dumps
- Still need to resolve a few things
- Should daily/nightly dumps be expanded beyond GeneACE? Sequence databases?
- Can we match WBPersons to ORCID IDs?
- We can ask WB users what their IDs are?
- Or, we can give WB users (once logged in) a way to link their WB account to an ORCID ID?
- How many WB authors already have an ORCID ID? We can search the ORCID database or contact WB authors?