To batch correct or analyse separately?

Question

[bompipi95]

Suppose I have a set of scRNA-seq datasets at hand that I want to analyse and run a typical scRNA-seq workflow.

What would the advantages/disadvantages be of applying batch-correction tools to integrate these datasets and analysing them jointly, vs analysing each dataset separately and then pooling the results together by some means?

Answer 1

[bompipi95]

Thanks both - found this quote from chapter 3 of the OSCA multisample chapter that addresses my question:

The greatest value of batch correction lies in facilitating cell-based analysis of population heterogeneity in a consistent manner across batches. Cluster 1 in batch A is the same as cluster 1 in batch B when the clustering is performed on the merged data. There is no need to identify mappings between separate clusterings, which might not even be possible when the clusters are not well-separated. By generating a single set of clusters for all batches, rather than requiring separate examination of each batch’s clusters, we avoid repeatedly paying the cost of manual interpretation. Another benefit is that the available number of cells is increased when all batches are combined, which allows for greater resolution of population structure in downstream analyses.

Answer 2

[ATpoint]

Instead of writing that all down I recommend to read the relevant part of the Bioconductor scRNA-seq book (OSCA) which covers the "how's" as well as the "why's" in great detail:

http://bioconductor.org/books/release/OSCA.multisample/integrating-datasets.html#motivation

[predeus]

Another great and up-to-date resource is this: https://www.sc-best-practices.org/preamble.html