Combining Conditions in scRNA-seq (Drop-seq) for Analysis

Question

[asyndeton17]

Hi,

I recently ran a Drop-seq experiment and got my demultiplexed data back for two different conditions. When should I combine the two files together? I began with two queryname-sorted BAM files. Should I merge the files after alignment, right before alignment, or after I extract the matrix for each condition?

Thanks

[Devon Ryan]

What's the goal? I assume you have multiple cells for each condition and are able to get counts for each. Normally you'll be using a matrix of that.

[asyndeton17]

Yes, there are many cells for each condition, but I want to plot both conditions onto one plot (probably a PCA plot).

Answer 1

[Devon Ryan]

For PCA (or tSNE, which is likely what you'll end up doing), you'll need a matrix with all of the cells for both conditions. Don't "merge" things in any other way.

[asyndeton17]

That seems reasonable. Will I be able to keep track of which condition cells came from within the matrix? If so, how?

[Devon Ryan]

Normally you just cbind() them and will know beforehand how many come from each cell type, so yes. But the responsibility for doing that is on you.