batch correction limma package (bead based assay data)

Question

[MC_J]

Hi,

I need to know if I can use limma package to remove batch effect, but my data is from a beads based assay.

Thanks!

Answer 1

[ATpoint]

Limma handles batch effects by including it into the design (so the model.matrix), please see the limma user guide over at Bioconductor which has a lot of code examples and case studies. https://www.bioconductor.org/packages/release/bioc/vignettes/limma/inst/doc/usersguide.pdf

[MC_J]

Many thanks! I’m planning to use Limma for batch correction, but I’ve been advised that the corrected data should generally be used only for visual analyses, like heatmaps, PCA, etc. When doing other types of analyses, it’s recommended to include batch as a variable in the model, if possible.

However, non-parametric tests like Kruskal-Wallis don’t allow for batch inclusion, whereas tests like ANOVA do such as when including variables like sex and age. I understand that Limma can transform data for use in parametric statistical tests like ANOVA. Could someone tell me which function in Limma I should use for this?

For context, my study is focused on comparing various antibodies between two groups with different levels of disease severity.

Thanks very much!

[ATpoint]

Read the user guide. It's discussed in there how to do exactly what you ask.