Return version of gnomAD and ExAC

[selinad]

Would it be possible to return the version of ExAC, 1000G and ESP on the page via myvariant.info? Curators would like to see this.

If not, we can add a noter that basically says the data is what is currently presented in these external resources (or something similar)

[wrightmw]

MyVariant.info show version numbers in http://myvariant.info/v1/metadata (e.g. ExAC is currently version 0.3.1)

[selinad]

@wrightmw @jimmyzhen would this be reasonable to do for R15 alongside gnomAD data?

[wrightmw]

Links to #1523

[wrightmw]

@selinad @jimmyzhen

MyVariant.info show version numbers in http://myvariant.info/v1/metadata

1. gnomAD version will be available with their next release (due early May).

2. The in silico predictors could just be included under the dbNSFP version number.

3. MyVariant.info includes ClinVar versioning

4. 1KG version will need to be obtained from Ensembl.... although they don't seem to have the version number in the file we use (http://rest.ensembl.org/variation/human/rs1801133?content-type=application/json;pops=1;population_genotypes=1). The current 1KG version is 2013050.

[selinad]

thx so much for all of this info, @wrightmw !

[selinad]

Also, when someone evaluates something with this data and we store it as a "blob," this information should also be stored in the blob

[jimmyzhen]

Hi @gcheung-SF,

I think this may be a good task for you to tackle for the next few days and to become more familiar with the VCI.

Please feel free to chat with me and/or @wrightmw about this. Thank you!

[jimmyzhen]

@gcheung-SF,

The blob mentioned above refers to a JSON object used to store the population data pertinent to a given variant when the user saves an evaluation on the Population tab in the VCI.

The JSON object is actually a linked object defined in the following schema: https://github.com/ClinGen/clincoded/blob/dev/src/clincoded/schemas/evaluation.json#L25-L30 https://github.com/ClinGen/clincoded/blob/dev/src/clincoded/schemas/population.json

Below is an example of what the JSON data of a typical evaluation looks like: https://curation-test.clinicalgenome.org/evaluations/294446fd-eeb2-4d98-8feb-dd635a62bc4d/?format=json

Note that the populationData object (the blob) in the tree. It consists of gnomAD, exac, esp and tGenomes children objects in which various population data is stored.

The requested changes in this ticket, if I am not misinterpreting, will be:

1. Display the versions of the external data sources in the UI (e.g. various table/panels in the Population tab in the VCI.
2. Store the versions along with each individual children objects in the populationData object.

@wrightmw, did I miss anything here?

[jimmyzhen]

Hi @gcheung-SF,

I briefly took a look at your code so far. Below are my thoughts:

1. The following metadata request and its subsequent data handling can be potentially decoupled from the fetchMyVariantInfo() function (e.g. https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/index.js#L227-L231). It can be a function of its own since it does not require the use of variant data.
2. The following setState seems unnecessary (e.g. https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/interpretation/population.js#L145, https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/interpretation/population.js#L196).
3. I am not sure whether this is preferable (e.g. https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/interpretation/population.js#L968-L970).
4. I probably prefer to change this line (e.g. https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/interpretation/population.js#L975) to something like: {version ? <span>Version: {version}</span> : null}
5. You may want to confirm with @wrightmw on this - does it make sense to display the version number in the event when there is no gnomAD and/or ExAC population data being returned?
6. You may want to coordinate with @markmandell on this - Mark is currently working on allowing gnomAD and ExAC population data to be displayed for Indel variants (in addition to single nucleotide). As a result, the singleNucleotide condition at https://github.com/ClinGen/clincoded/blob/7f9b77e385657bd490034751e6ac374b25e1eb31/src/clincoded/static/components/variant_central/interpretation/population.js#L961 can be problematic for rendering the version if the variant in question happens to be a deletion or insertion.

Please feel free to let me know if you have any questions on the above.

[wrightmw]

@jimmyzhen @gcheung-SF In answer to point 5. Yes, it does make sense to show the version number even if there is no data returned. In essence we are stating that in that specific version number of gnomAD and/or ExAC there is no allele frequency data for that variant. In future versions there may be data for that variant.

[wrightmw]

Once we show versioning, we should add a note reminding people to check whether there is a newer version available at the orginal source (e.g gnomAD).

[gcheung-SF]

@wrightmw , I can work on adding that if you can show me where to add it. Thanks!

[gcheung-SF]

@wrightmw , My change to display the version on the population tab is ready for review. Please try it out on https://gctest.demo.clinicalgenome.org and let me know if there's any issues. Thanks!

[wrightmw]

Thanks @gcheung-SF ...looks great

[gcheung-SF]

@bryanwulf, Please review the PR - https://github.com/ClinGen/clincoded/pull/1860. Test site - https://gctest.demo.clinicalgenome.org/variant-central/?variant=ca55d9fe-3ccf-40b1-ae25-bb63fc5d23a3&tab=population

Thanks.

[bryanwulf]

@wrightmw It looks like version information doesn't appear when there's no population data: https://gctest.demo.clinicalgenome.org/variant-central/?variant=79444ab6-d275-4b33-ab44-5ce82bfbbc81&interpretation=70dd3912-b6cd-4f6a-8b53-765329c3fec6&tab=population

Does this need to be fixed for this release?

[cgpreston]

Tested with 139214: See gnomAD/ExAC version listed (as expected, variant in both). Tested with 41231, see gnomAD/ExAC version listed (as expected, variant NOT in either) Tested with 25342, see gnomAD/ExAC version listed (as expected, variant in gnomAD, not in ExAC) Tested with CA352991, see gnomAD/ExaC version listed (as expected, variant too large to return data for either) -

[wrightmw]

In R25 release

[wrightmw]

R29 testing: Looks great!

[wrightmw]

In R29 release.