Handling limma usage after running multiple imputation on the covariates with missing data
Entering edit mode
S • 0
Last seen 7 weeks ago

I have missing data on my covariates of interest in my samples (a human longitudinal study) on which I am going to run limma.

Multiple imputations in R give multiple datasets. Is it advisable to run limma using each of these imputed datasets and then average the estimates somehow that I get from limma or is there a better way of doing this that I am missing out on?

The missingness is between 10 to 13% in these covariates.

limma • 149 views
Entering edit mode
Last seen 43 minutes ago
WEHI, Melbourne, Australia

limma is not designed to be used with multiply imputed covariates, so I can't offer a solution that uses the multiple imputes.

Averaging the results from multiple runs defeats the purpose of multiple imputation. If averaging was your plan, then it would be better to use mean imputation in the first place instead of multiple imputation. With multiple imputation, you could run limma on a few different imputations and see if the results remain stable. If they do, then you can use any of the results.


Login before adding your answer.

Traffic: 203 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6