about the batch effects

Question

[Bogdan]

Dear all,

if I may ask for your opinions, and to double-check please : shall we have 2 RNA-seq samples

(sample1, sample2), and :

-- the same/ or distinct time points of COLLECTING the CELLS (of sample1, sample2), or

-- the same/ or distinct time points of SEQUENCING (of sample1, sample2),

in all these situations below, we shall correct for the BATCH EFFECTS, correct ? thank you !

SITUATION1 :

-- sample1: collection time A1, sequencing time B

-- sample2: collection time A2, sequencing time B

SITUATION2 :

-- sample1: collection time A, sequencing time B1

-- sample2: collection time A, sequencing time B2

SITUATION3 :

-- sample1: collection time A1, sequencing time B1

-- sample2: collection time A2, sequencing time B2

[ashish]

Did you first check if there are any batch effects in the 3 situations?

[Bogdan]

hi Ashish.

On the actual data that we have, we do see big differences between the replicates on the PCA plot.

my question was a bit more general, in order to develop a better approach.

At this moment, 'm setting up a matrix, where we do have : SAMPLE; COLLECTION_TIME; SEQUENCING_TIME, planning to use the linear models in LIMMA.

Answer 1

[swbarnes2]

In general, running the same library at different times is not going to cause a batch effect. So don't worry about when you sequence. Do collect all the RNA at around the same time, and do make all the library preps at the same time, as much as you are able.