Question: Can differential gene expression heatmaps be compared across multiple separate experiments?
0
gravatar for NHEJ
5.6 years ago by
NHEJ320
United States
NHEJ320 wrote:

Can heatmaps be compared across multiple different experiments?  Obviously, you can compare sample runs within a single heatmap, but what about comparing sample runs in one heatmap to sample runs in another heatmap (e.g., if I want to track the activity of a gene across multiple therapeutic treatments which I performed in separate experiments from last year)?  I am worried that I may run into statistical issues when doing this kind of cross-comparison, since as far as I'm aware the heatmap and heatmap.2 commands in R perform the proper coloring of the heatmaps within a single dataset (not multiple datasets).

heatmap rna-seq • 3.1k views
ADD COMMENTlink modified 5.6 years ago by Jason900 • written 5.6 years ago by NHEJ320
0
gravatar for Jason
5.6 years ago by
Jason900
United States
Jason900 wrote:

I would merge the data into the same heatmap. If you have a sample that was included in both the RNA-seq from a year ago and the more recent run you should compare those first to see what the batch to batch variation is like. If they are very similar (R^2 > 0.9) then maybe you can just go ahead just try merging all the samples in the same heatmap.  If there's significant variation or you don't have a common replicate you may try normalizing or scaling all the data to fall within the same range of values then generating the heatmap. 

ADD COMMENTlink written 5.6 years ago by Jason900

I understand that even though merging is probably the safer option, is it still safe to compare separate heat maps (given that they share common replicates)?

ADD REPLYlink written 5.6 years ago by NHEJ320

Maybe if you made sure the colors were exactly the same (i.e. -1 log2FC = green and +1 = red) it would be at least relatively interpretable in a qualitative sense. Without merging them all together it might be difficult to determine if the same genes are illustrating the same patterns. 

ADD REPLYlink written 5.6 years ago by Jason900

Realistically speaking, the heatmaps are just for diagnostics anyway, you're not going to get any real results from them. If you want to make real comparisons then you'll need to reanalyse things in a single go.

ADD REPLYlink written 5.6 years ago by Devon Ryan96k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1533 users visited in the last hour