COMETS 1.4 Pairwise Analysis Exposure UID

[nateappel]

In the pairwise analysis, the exposure_uid and exposure columns are reporting the exposurespec. Would it be possible to get the actual UID and name in these two columns.

pairwise_ouput_sample.xlsx

[ellatemprosa]

i can look into this since i was trying to fix exposure spec

[steven-moore]

When "All metabolites" is the outcome, meta-data are not being pulled in

[steven-moore]

Hi Dave,

Could you clarify?

S

[daveruggieri]

Sure. Rather than populating the exposure_uid and exposure fields with UIDs and biochemical names, they're currently being populated by "exposure_spec". I think this bug is much easier to visually conceptualize than to read, so I'm attaching an example. I used highlighting to distinguish between "spec" (blue), "UID" (green), and "metabolite name" (yellow). The problem cells are N11:O15, and their expected values are N19:O23.

Let me know if you have questions. comets.pairwise.results.exposure.example.xlsx

EDIT: I just realized the header in the 2nd grouping still says Gender Adjusted Models. That's actually an example of the Pairwise output.

[steven-moore]

If I recall correctly, a key concern here was that the outcome name was being manipulated before being output. This made it challenging to correctly link to exposures for analyses, e.g. preparing a metabolite*metabolite heat-map, meta-analysis, etc. This has largely been mitigated with Ewy's edits--exposure and outcome names are being passed through without any manipulation.

Dave is nonetheless correct that the "exposure_uid" values are incorrect (see his text and example above). In general, I want to make sure that "exposure" and "outcome" are being treated identically in processing; any customized treatment of exposure vs. outcome has the potential to create problems.

[steven-moore]

Per our phone conversation today, Ella will look into this and issue #22 (the two issues are interrelated)

[steven-moore]

The example above was somewhat dated, so I have included a new example based on examining the first three metabolites, in alphabetical order, correlated with one another. In the screenshot, you can see that the metabolite names were swapped for the "outcomes' but not the "exposures"

In addition, this can also be seen in the download file:

[steven-moore]

Fixed and closed