January, 2020 Community Call - Kidney
Presenters:
Anna Greka - Broad Institute
Becky Steck - Michigan Medicine
Evren Azeloglu - Icahn School of Medicine at Mount Sinai
Slides:
Video recording:
Links:
- Kidney Cell Atlas
- Kidney Precision Medicine Project (KPMP)
- KPMP github
- Kidney Tissue Atlas Ontology (KTAO)
Summary of the January
Key learnings
- Upcoming specific kidney area focus and scATACSeq, slide-seq, HCR/FISH, metabolomics, proteomics
- Harmonising metadata, vocabulary & ontologies between initiatives (KPMP, HubMap, HCA) will help build the atlas
- Strength of engaging directly with the clinicians, pathologists and researchers to address gaps, inconsistencies and errors in the ontologies
- Modular approach to metadata structure design works well for the different data types in the project but had to de-modularise in order to simplify data re-use
- Difficult to use the specifications defined in a spreadsheet to then validate incoming metadata
- Important to engage all points of metadata collection early in the process
Needs from the community
- Way of harmonising metadata across consortia.
- Need to understand important metadata fields from scATACSeq, slide-seq, HCR/FISH.
- How to engage with reviewing and updating key ontologies such as UBERON and CL
- Tools for how to define metadata schema
- Generic validation tools that can be used with per consortium metadata schema
Ideas for how to action
- Given it is unlikely that we can conform to one united standard, can we build tools that allow interoperability/translation across all the standards? Are there fields that are “non-negotiable” across all the consortia?
- Can we have a standard interoperability format that would allow metadata translation?
- How do we ease the burden of how to map between standards? Are automated tools possible?
- Improve communication with the ontology team or act as an intermediary between researchers/consortia and SPOT?
- More approachable web resources on how to provide feedback on ontologies
- Can our json validator be used by anyone with schema as json? Is there adequate documentation to enable others to use it?