True meaning of global transcriptome analysis

Question

[babasaraki]

Dear All,

I have two different cancer cell lines (EJ28 and TCCSUP) that are persistently infected with virus and called them EJ28Pi and TCCSUPPi respectively. So, I performed RNA-seq analysis for each cell line versus its control (for example EJ28Pi VS EJ28) differently using DESEq2 package from R.

Now I want to perform global expression between the two cell lines. The question is how can I do this?, am I to combine the triplicate of both EJ28Pi VS EJ28 and TCCSUPPi VS TCCSUP together in a single analysis or I should just combine the triplicate of EJ28Pi VS TCCSUPPi? I am a bit confused here. So, I need the true meaning and brief explanation of Global Transcriptome Analysis in my case using my samples above.

Thank you for your usual support.

[Kevin Blighe]

"Global Transcriptome Analysis" <- from where have you heard this term, and in which context? It explains very little about what you actually [may] want to do. So, please tell us what you actually want to do.

[babasaraki]

Thank you Dr. Kevin for your prompt response. I have heard this term in so many different papers including nature articles. Refer to the following links frontier publishers, ASM Journal of Bacteriology in ASM and the nature publisher Scientific Reports. In my case, I want to see the changes in gene expression due to persistent viral infection from the established persistently infected cell lines (EJ28Pi and TCCSUPPi). So, I initially analysed the cells differently e.g. EJ28Pi VS EJ28 (control), and TCCSUPPi VS TCCSUP(control). However, now I want to know in which context does the term global expression or global transcriptome expression is applied. Is it by comparing the transcripts expression of EJ28Pi versus TCCSUPPi or is by comparing the triplicate of both EJ28Pi VS EJ28 and TCCSUPPi VS TCCSUP together in a single analysis?

Thank you.

[Kevin Blighe]

No problem, Professor babasaraki. Oh, I see, but they are just using that term in a very general way, in the manuscript title. It just means that they have data from the [assumed] entire / global transcriptome, which they used for their analyses. So, 'global transcriptome analysis' can relate to any type of analysis that is performed using data from the transcriptome.

You may take the advice from Ahill in their answer (see below).

[babasaraki]

Thank you Sir. I appreciated. I am just a PhD student not to talk of a Professorial cadre. Just want to get the meaning of what I am doing in the analysis.

[Kevin Blighe]

Many PhD students have more talent and maturity than Full Professors. You will be fine.

[babasaraki]

Thank you dear Professor for the compliment and words of motivation.

Answer 1

[Ahill]

I'm assuming the triplicates you mention mean that you have n=3 of each of the 4 experimental conditions, for a total of 12 samples. If you want to compare cell lines, unless infection has negligible effects, you do not want to combine infected and non-infected cells into a single group. Instead, you could fit a statistical model to all n=12 of your samples of the form expr ~ cell_line + infection + cell_line:infection, to identify transcripts that are are different between the cell lines, affected by infection, or influenced by an interaction between cell line and infection.

[babasaraki]

Exactly you got the point. I did analysed both the 12 samples at the same time with their replicate but I am not sure whether is appropriate to do that by comparing all the 12 samples or just the infected cell lines (EJ28Pi VS TCCSUPPi). When is the global transcriptome is applied in my context. Could it be when combinining all the 12 samples in the analysis or when combining only the infected cell lines?

Thank you.