EPIC GWAS IMPORT ERROR #50

Merged
merged 9 commits into from Jun 1, 2012

Conversation

Projects
None yet
3 participants
@paddie
Contributor

paddie commented May 31, 2012

I've been having fevery dreams these last couple of weeks, and I suddenly realized why. The import used one idiotic assumption when loading the GWAS. That each study only references a single disease. This, in retrospect, is obviously not the case, given that multiple GWAS correlations could possibly be identified in a study.

This means that each study now has a list of diseases they reference, instead of just one, and each disease now has many more studies to reference (possibly).

In other words, there is actually a many-to-many relationship between disease and study, whereas the old model only had a one disease to many studies relation.

  • added disease <----> study relation
  • updated templates to reflect this model change
  • study no longer has a study.disease_ref but an array of keys

Since this fix obviously has implications on how we search for studies, I wanted to give @wejendorp a heads up. Since the "disease_trait" field is no longer there, replaced by an array of disease_keys (I could add an additional disease_names string-array if that would simplify things?), you can no longer index the disease-field of the study. For now I've simply commented out that field in your indexing, and I couldn't figure out how to add indexing for arrays.

Patrick-Ranjit D. Madsen added some commits May 30, 2012

Patrick-Ranjit D. Madsen
Some simple methods to handle the many-to-many relation
* disease <*------>* study because some studies apparently identify multiple disease relations
Patrick-Ranjit D. Madsen
Epic fix in GWAS imports
- that's right, ONE STUDY CAN REFERENCE MULTIPLE DISEASES
@paddie

This comment has been minimized.

Show comment Hide comment
@paddie

paddie May 31, 2012

Contributor

example url with an example of a study with multiple identified GWAS relations <localhost>/study/19010793

remember to purge() and populate() first

Contributor

paddie commented May 31, 2012

example url with an example of a study with multiple identified GWAS relations <localhost>/study/19010793

remember to purge() and populate() first

@ghost ghost assigned wejendorp and jensraaby May 31, 2012

@wejendorp

This comment has been minimized.

Show comment Hide comment
@wejendorp

wejendorp Jun 1, 2012

Contributor

Now the Study documents have a list of names of diseases (csv), and are fully searchable again.

Contributor

wejendorp commented Jun 1, 2012

Now the Study documents have a list of names of diseases (csv), and are fully searchable again.

@jensraaby

This comment has been minimized.

Show comment Hide comment
@jensraaby

jensraaby Jun 1, 2012

Contributor

Are we ready to close this one yet? Thinking we should be moving this off the production line so we can start on to the pagination story

Contributor

jensraaby commented Jun 1, 2012

Are we ready to close this one yet? Thinking we should be moving this off the production line so we can start on to the pagination story

@paddie

This comment has been minimized.

Show comment Hide comment
@paddie

paddie Jun 1, 2012

Contributor

Agreed, just want to be sure Jacob understands what happens to the indexer

Hilsen

Patrick-Ranjit D. Madsen

Sent from my mobile. Please excuse any and all typos, errors,
formatting, lack of punctuation or short responses.

On 01/06/2012, at 20.20, jensraaby
reply@reply.github.com
wrote:

Are we ready to close this one yet? Thinking we should be moving this off the production line so we can start on to the pagination story


Reply to this email directly or view it on GitHub:
#50 (comment)

Contributor

paddie commented Jun 1, 2012

Agreed, just want to be sure Jacob understands what happens to the indexer

Hilsen

Patrick-Ranjit D. Madsen

Sent from my mobile. Please excuse any and all typos, errors,
formatting, lack of punctuation or short responses.

On 01/06/2012, at 20.20, jensraaby
reply@reply.github.com
wrote:

Are we ready to close this one yet? Thinking we should be moving this off the production line so we can start on to the pagination story


Reply to this email directly or view it on GitHub:
#50 (comment)

@wejendorp

This comment has been minimized.

Show comment Hide comment
@wejendorp

wejendorp Jun 1, 2012

Contributor

Merge it! I have it under wraps ^^

Contributor

wejendorp commented Jun 1, 2012

Merge it! I have it under wraps ^^

wejendorp added a commit that referenced this pull request Jun 1, 2012

@wejendorp wejendorp merged commit 23e9a09 into master Jun 1, 2012

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment