extra column in evidence page: PheWAS and GWAS

ElaineMcA commented 6 years ago

@deniseOme commented on Mon Jul 24 2017

@elipapa has proposed a couple of changes to the genetic evidence table based on the new PheWAS data e.g.

https://phewas-catalog-test--paratrooper-antelope-70721.netlify.com/evidence/ENSG00000146648/EFO_0003952

We need to provide the summary statistics (e.g. cases, odds ratio). This info is not available through the PheWAS website. So linking the evidence page to the original source is not an option, @nikiforosk @iandunham @jcbarret.

A new column has been proposed by @elipapa @emepyc that would contain the addition info, which is evidence-specific. A screenshot of this will be available soon.

The questions are:

How will we populate the same column for GWAS? Will we leave the column empty? Yes, we could and explain the reason if/when users complain.

Or will we provide these stats as well for GWAS? Yes, we could. Users have already asked us in the past to provide odds ratio and confidence intervals for GWAS. At the time we pointed them to Ensembl where they have the details.

If we decide to provide OR, CI or/and beta-coefficient for GWAS to be consistent with what is provided for PheWAS, @ckongEbi would have to liaise with Olga et al, as this info is not given to us yet.

@iandunham commented on Mon Jul 24 2017

I think it may be best to have a separate table for the Phewas if
  we really want the cases. The odds ratio we can get for GWAS
  traits but the cases for the Phewas data has a specific meaning
  e.g. 25 means that 25 of the 13,835 patients in the Phewas study
    had the phenotype. In the GWAS studies we would generally have a
    specific number of patients and matched controls. So putting the
    two together will be misleading.
In both cases
    the sample size (cases) is reflected in the potential values
    that the OR and p value can take, so it might be better to have
    the OR and 95% CI as this avoids the user having to work out
    what effect the case number would have other than hand waving
    saying it's a small number of cases. These are available in
    Supplementary Table 8 of the paper. Having these and not the
    cases would enable both data to readily fit in the same table.

For GWAS we do
    use the population size to assess whether the study is reliable,
    but this is encapsulated in the score in a relatively low
    resolution way. Ask Olga for the OR and CIs, but we will also
    need to make sure we get these for the other GWAS sources via
    postgap.

On 24/07/2017 11:10, deniseOme wrote:

  @elipapa has proposed a couple of
    changes to the genetic evidence table based on the new PheWAS
    data e.g.
  https://phewas-catalog-test--paratrooper-antelope-70721.netlify.com/evidence/ENSG00000146648/EFO_0003952
  We need to provide the summary statistics (e.g. cases, odds
    ratio). This info is not available through the PheWAS website.
    So linking the evidence page to the original source is not an
    option, @nikiforosk @iandunham @jcbarret.
  A new column has been proposed by @elipapa @emepyc that would contain the
    addition info, which is evidence-specific. A screenshot of this
    will be available soon.
  The questions are:
  How will we populate the same column for GWAS?
    Will we leave the column empty?
    Yes, we could and explain the reason if/when users complain.
  Or will we provide these stats as well for GWAS?
    Yes, we could. Users have already asked us in the past to
    provide odds ratio and confidence intervals for GWAS. At the
    time we pointed them to Ensembl where they have the details.
  If we decide to provide OR, CI or/and beta-coefficient for GWAS
    to be consistent with what is provided for PheWAS, @ckongEbi would have to liaise with
    Olga et al, as this info is not given to us yet.
  —
    You are receiving this because you were mentioned.
    Reply to this email directly, view it on GitHub, or mute the thread.

  {"api_version":"1.0","publisher":{"api_key":"05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":{"external_key":"github/opentargets/webapp","title":"opentargets/webapp","subtitle":"GitHub repository","main_image_url":"https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://cloud.githubusercontent.com/assets/143418/15842166/7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in GitHub","url":"https://github.com/opentargets/webapp"}},"updates":{"snippets":[{"icon":"DESCRIPTION","message":"extra column in evidence page: PheWAS and GWAS (#164)"}],"action":{"name":"View Issue","url":"https://github.com/opentargets/webapp/issues/164"}}}

@andrewhercules commented on Fri Aug 10 2018

Triage notes:

Related to issue #141 in opentargets/roadmap repo
Have identified this as a suitable feature for a mini-release
Will require creating a new column in the Common Diseases table and exposing data.unique_association_fields.odd_ratio and data.unique_association_fields.confidence_interval if available
Data already available from OT API - see sample API response for NOD2 and Crohn's disease

LucaFumis commented 6 years ago

We've tried 2 options: 1) display odds ratio + confidence interval + cases in one column 2) add 3 cols and display odds ratio, confidence interval, cases in a column each issue54_opt1 issue54_opt2

andrewhercules commented 6 years ago

Following up from yesterday's meeting, it was decided that the table will be like the first screenshot, with a combined column Odds Ratio (Confidence Interval) and the sorting will be done on the odds ratio value.

We can also hide the number of cases as that's available via the API.

LucaFumis commented 6 years ago

Deployed in version 3.7.4

opentargets / issues

extra column in evidence page: PheWAS and GWAS #28