Time course analysis of RNAseq data

Question

[nicolas.hipp]

Hello everyone,

I have a time course dataset from RNAsequencing. Tipically it's bulk RNAseq, with 8 time-points and 3 replicates per sample. I was wondering if there is something like the "pseudo-time" from monocle package (single cell RNAseq) to order my sample on one dimension? In the PCA generated with Deseq2 plotPCA(rlog (DESeqDataSetFromMatrix), it seem to have a continuum between the sample in the axe 1. It's a differentiation process from primary cells. Problem is, as it's primary cells, activation between replicate is heterogeneous and I get for example Sample1-Day3 before Sample2-Day2.

So I would like to order sample without any a priori in one dimension. Do you know if there is a soft that can do that? Thanks a lot, Nicolas

Answer 1

[Jean-Karim Heriche]

I can think of two approaches. One would be to use dynamic time warping to reassign time points, the other would be to use seriation/ordination (see also the documentation of the R package seriation).

[nicolas.hipp]

Hi, Thanks for these very informative informations, It's seem to really fit with my question ;)