Mis-matched correlations

[kovasap]

Inputs.csv Biomark.csv

"I noticed a few bugs when looking at correlations with CRP in the attached files. For example, B1's p-value is 0.58 in my data but 0.47 in the correlator. B2's correlation with CRP in my data has a p-value of 0.34, 0.29 in the correlator. Some others are off by 0.01 or 0.02, which isn't much different, though."

[kovasap]

Strangely, after importing this test case and running it in isolation, I get even more variable results (correlation of 0.132 between CRP and B1)!

[kovasap]

According to Google Sheets, 0.132 is in fact the correct value. So the real issue must lie in how the data is being brought to this point. Interestingly this suggests that @MichaelLustgarten your method is also wrong potentially?

[ghost]

Thanks Kovas. I'm using SAS Enterprise Guide, which is point and click, with the same dataset. Also, let's switch to gmail...

Michael Lustgarten, PhD Scientist II Nutrition, Exercise Physiology, and Sarcopenia Laboratory Jean Mayer Human Nutrition Research Center on Aging Tufts University Boston MA 02111

---

From: Kovas Palunas @.> Sent: Wednesday, March 23, 2022 9:53 PM To: kovasap/biomarker-correlator @.> Cc: Lustgarten, Michael @.>; Mention @.> Subject: Re: [kovasap/biomarker-correlator] Mis-matched correlations (Issue #2)

According to Google Sheets, 0.132 is in fact the correct value: https://docs.google.com/spreadsheets/d/17ysvkCHZcZBlx_tn0ud9fJxD8P-pu37bTvPvAWzkzdM/edit#gid=0. So the real issue must lie in how the data is being brought to this point. Interestingly this suggests that @MichaelLustgartenhttps://github.com/MichaelLustgarten your method is also wrong potentially?

— Reply to this email directly, view it on GitHubhttps://github.com/kovasap/biomarker-correlator/issues/2#issuecomment-1076988172, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AYCKOQZLZUP2XNAPIWD74WDVBPDJZANCNFSM5RPV4MPA. You are receiving this because you were mentioned.Message ID: @.***>

[kovasap]

Got it, added your gmail as a collaborator on github. You could also change your github email to your gmail if you don't want another account created.

On Thu, Mar 24, 2022 at 3:06 AM MichaelLustgarten @.***> wrote:

Thanks Kovas. I'm using SAS Enterprise Guide, which is point and click, with the same dataset. Also, let's switch to gmail...

Michael Lustgarten, PhD Scientist II Nutrition, Exercise Physiology, and Sarcopenia Laboratory Jean Mayer Human Nutrition Research Center on Aging Tufts University Boston MA 02111

---

From: Kovas Palunas @.> Sent: Wednesday, March 23, 2022 9:53 PM To: kovasap/biomarker-correlator @.> Cc: Lustgarten, Michael @.>; Mention @.> Subject: Re: [kovasap/biomarker-correlator] Mis-matched correlations (Issue #2)

According to Google Sheets, 0.132 is in fact the correct value: https://docs.google.com/spreadsheets/d/17ysvkCHZcZBlx_tn0ud9fJxD8P-pu37bTvPvAWzkzdM/edit#gid=0. So the real issue must lie in how the data is being brought to this point. Interestingly this suggests that @MichaelLustgarten< https://github.com/MichaelLustgarten> your method is also wrong potentially?

— Reply to this email directly, view it on GitHub< https://github.com/kovasap/biomarker-correlator/issues/2#issuecomment-1076988172>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AYCKOQZLZUP2XNAPIWD74WDVBPDJZANCNFSM5RPV4MPA

. You are receiving this because you were mentioned.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/kovasap/biomarker-correlator/issues/2#issuecomment-1077453430, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACC3NASHQATZLOZDDFZVATDVBQ5EBANCNFSM5RPV4MPA . You are receiving this because you authored the thread.Message ID: @.***>

[kovasap]

I can't seem to reproduce your problem. I'm always getting 0.132 with whatever method i use. Here is my attempt at using the files linked in the first post of this issue: