WormBase-Caltech Weekly Calls September 2016

September 1, 2016

Working group discussions

• More clear about issues surrounding orthology sets
• ZFIN has protein curation that may be becoming a bit out of date

Single-cell RNA profiling

• We've been curating some data, but not clear how to display the data
• Some data sets, like tiling arrays, might conclude genes that are expressed in a cell/tissue
• Genes that are detected at convincing levels (present call) in cell/tissue
• Should we display expression data that we do not (or cannot) update differently from other data
• Simply: is a gene in a cell/tissue or not; is it enriched in a cell/tissue or not
• Could capture quantitative assessments; transcripts per million, etc.
• Should we provide an analysis of raw data to users to perform their own comparative analyses
• Bob Goldstein's group prepared software tools in their paper; could we make use of them?
• Big question: how do we handle divergent (expression) data?
• David (Angeles) working on an expression data analysis pipeline
• Single-cell RNA-Seq data: should it be merged with expression pattern data? Ideally, yes
• There's difficulty in reconciling small and large datasets (or different data types)
• Would we want to apply some type of weighting to each type of evidence for a gene-anatomy association, for example
• In general for expression, it would be great to consolidate data to directly connect genes to anatomy/life stage

September 8, 2016

TransgeneOme import (Daniela)

• adding a Community_curator tag or Curator tag in the Expression model to acknowledge the community curators that annotated/will annotate the data.
• for the records- documentation on the import here: http://wiki.wormbase.org/index.php/TransgeneOme_import
• Daniela and Juancarlos worked on parsing large JSON import
• Data is dynamic: people can add images and annotations; will be updated every WormBase release
• Do we want to import the person that created the annotation via the TransgeneOme project?
  • Maybe, but should be free text instead of a modeled tag for WBPersons (Raymond)
  • It is likely sufficient to point WB users to the TransgeneOme annotation, without explicitly referencing the contributor

Phenotype assay model proposal (Chris)

• Trying to create a model that can accommodate complex genotypes and make querying phenotype experimental data easier
• Want to co-opt an existing, unused model ?Phenotype_assay to encapsulate details of an experiment
• Proposal is to pull several tags from ?RNAi and #Phenotype_info into the ?Phenotype_assay class
• Remaining tags in ?RNAi could be used to hold RNAi mapping information
• Remaining rags in #Phenotype_info can be used to describe the nature of an allele in the context of a phenotype, or nature of phenotype (Ease of scoring)
• Maybe should keep penetrance, quality and EQ annotations in #Phenotype_info so as to provide detail for each phenotype when multiple phenotypes are used
• It is currently difficult to recreate a phenotype annotation made by a curator using the OA, as so much experimental detail is captured exclusively in the hash

September 15, 2016

Strain curation with disease

• People create transgenic strains for modeling human disease
• We currently only curate strains in CGC (where we get our strain list)
• Other strains have been added via direct communication with Mary Ann
• Ranjana would like to capture strains affecting human disease from the literature
• Many strains not available in WormBase
• Is it worth curating these other strains? Yes; one concern that we would be diluting the quality of existing strains
• Big concern is whether or not the strain(s) are available
• (Raymond) Two aspects of strain curation: 1) strains that are available (e.g. from CGC) 2) carrier of genotype to use for other curation (like disease)
• (Raymond) We could just capture genotype as free text
• Would be best to curate the strains, regardless of availability
• Ranjana will send some example papers/strains to Mary Ann and they will discuss
• We could create links from strains in WB to the CGC (or state that people should write directly to the source lab; might be problematic).

WormBase header changes

• Miyuki has been working on new WormBase header; requests feedback
• Available on staging

Tissue enrichment analysis

• Tool is live and ready to use
• Raymond has added a link to the tool on the Tools page
• Raymond will submit a ticket to put on the main Tools dropdown

September 22, 2016

GOC meeting at USC, November 4-6, Noctua Workshop, November 7

• Next GO Consortium meeting is in early November at USC
• The consortium meeting will run from Friday, 11/4 - Sunday, 11/6
• Noctua/LEGO workshop will be on Monday, 11/7 (9am - 3pm)
• Meeting Logistics Page
  • We will need to add all WB curators to the attendees table
• Meeting Agenda - in progress
• People interested in going to the consortium meeting or the workshop should register or let Kimberly know
• Kimberly can setup accounts for people to try Noctua trial curation

Display of top community curators

• Sibyl has setup a widget for the homepage and submit data page displaying top 20 community curators
• Chris has compiled numbers from phenotype community annotations
• We want to gather community annotations from other forms as well (historically and from newer forms)

Intellectual Lineage

• Will display on person page
• There exist "unknown" relationships; can we update those to known relationships?
• We can email both parties of each unknown relationship asking about the nature of the relationship

Gene Ontology development

• There may be a need for an additional ontology developer/editor
• There's an accumulation of GitHub tickets requesting new terms/edits
• It's not clear how the ontology will evolve in light of LEGO curation

Tables in PDF

• Some tables have been converted from Excel to PDF
• PDF tables are difficult to parse
• We can/should ask authors for original Excel spreadhseet

Metabolomics efforts

• PIs working on grants to study metabolomics in accumulated mutation lines
• Want to contribute data to WormCyc
• BioCyc doesn't use same standard vocabularies that WormBase does (e.g. ChEBI)

AGR Portal use cases

• Working group had started working on a gene page, but have made a new use case for general search
• The draft use case would describe searching with general text to find matches to genes, diseases, GO terms, and orthology groups
• Next will try to work out a use case for arriving at a gene comparison page, e.g. starting with a human gene
• Would probably be good to break into smaller groups of 2-3 people to independently work on different use cases
• Starting simple is good, even if the added value isn't immediately clear

September 29, 2016

AGR Portal Use Case update

• Use case working group now working on a gene/gene-product comparison page for all genes in an orthology group/PANTHER family or subfamily
• Plan is to display information for genes side-by-side for each data type
• First data types would include gene descriptions, gene ontology annotations, and disease associations
• For GO terms many would like to see the "ribbon" display with GO slim terms or context-specific terms
• Would be good if ribbon content was dynamically generated (rather than a predetermined set of terms), to reflect the nature of genes in an orthology group
• Could be useful to use SObA logic in selecting terms to show in ribbon
• Would be nice to display the genes according to species or listed by their relationships within a phylogenetic tree, for each data type
• Highlighting commonalities between genes in the group would be nice

Intellectual lineage display

• Juancarlos working on display of the lineage on WBPerson pages in a widget
• Data will be pulled from ACEDB and used to generate display
• Performing necessary calculations was crashing the machine
• Want to scale the display based on number of connections

Changes to Disease Term OA

• This OA will be used when information about the disease model cannot be curated to a C. elegans gene, e.g. when a human gene was put into the worm.
• Changes incorporate information for strains, genotypes, treatments like drugs etc.
• There's still a question about how to encapsulate the information from a single annotation
• Separate annotations will be made if a second paper uses the same disease model (described in the paper being curated) and adds a treatment to it to show amelioration or exacerbation of disease; separate annotation means a different row with a different PGID in the OA.
• Modeling can be done once more trial curation is done, some AGR modeling decisions to be made
• Could curate and store data in OA for now, holding it until we have an appropriate data model and display pipeline
• Want to capture strain, drug treatment, conditions, etc.
• How do we connect disease-relevant genetic interactions to the disease or disease models?
• Disease group working out standards for disease model curation
• Ranjana will email interaction curators when she needs to, in these cases

Noctua/LEGO

• Can we annotate details of experiment in Noctua? It's not clear
• Noctua has predominantly been used in the context of GO annotations
• Kimberly happy to demo Noctua in group meeting or individual Skype session