search

thomazbastiaanssen / kronos

7 stars 1 forks source link

How does kronos handle repeated sampling from the same mouse? #4

Open EOMAK91 opened 2 days ago

EOMAK91 commented 2 days ago

Hi,

I had question about how kronos handles repeated sampling? I'm not familiar with circadian statistical modeling, but when we use mixed effects linear regression modeling, it is typical for you to specify the random-effects term, such as "Biological_Replicate_iD". In the kronos tutorial, Animal_ID is not unique, but rather a sample collected at a given timepoint. Do we need to consider repeated sampling for modeling circadian rhythms?

E

thomazbastiaanssen commented 1 day ago

Hi,

Thanks for your interest in kronos! Indeed, for mixed modelling I'd also use your approach. Currently, Kronos doesn't support mixed effects modelling - our initial studies didn't need at we didn't have repeated measures there - though I plan to expand kronos to support mixed models in the not too distant future.

For now, I would suggest scaling your data by scaling your outcome variables of interest by your 'Biological_Replicate_iD' variable (i.e., subtract the ID-specific mean) - this would be analogous to adding a random intercept, or, in lme4 syntax, + (1|Biological_Replicate_iD).

Best wishes,

Thomaz

EOMAK91 commented 1 day ago

Thank you for the prompt response Thomaz.

Would you recommend this approach for both omic dataframes and single variable data?

E