Accounting for technical batch effects

Differential Expression Analysis with limma in R

John Blischak

Instructor

What are technical batch effects?

Every batch of an experiment is slightly different
Need to balance variables of interest across batches
If properly balanced, batch effects can be removed

Diagnosing technical batch effects

Dimension reduction techniques:
- Principal Components Analysis (PCA)
- MultiDimensional Scaling (MDS)
Identify the largest sources of variation in a data set
Are the largest sources of variation correlated with the variables of interest or technical batch effects?

plotMDS

library(limma)
plotMDS(eset, labels = pData(eset)[, "time"], 
        gene.selection = "common")

removeBatchEffect

exprs(eset) <- removeBatchEffect(eset, 
                  batch = pData(eset)[,"batch"],
                  covariates = pData(eset)[,"rin"])

plotMDS(eset, labels = pData(eset)[, "time"], 
        gene.selection = "common")

Olfactory stem cells

7 treatments, 4 batches
- Bioconductor package: HarmanData, Harman
- Osmond-McLeod et al. 2013, Oytam et al. 2016

table(pData(eset))

         batch
treatment b1 b2 b3 b4
       t1  1  1  1  1
       t2  1  1  1  1
       t3  1  1  1  1
       t4  1  1  1  1
       t5  1  1  1  1
       t6  1  1  1  1
       t7  1  1  1  1

Let's practice!

Differential Expression Analysis with limma in R