How can I compare the similarities of two transcriptomes?

Question

[avelarbio46]

Hello everyone! We have created a dataset consisting of different cells with a whole-transcriptome (50M reads). Let's say I have a transcriptome of one expanded cancerous cell, and also I have the transcriptome of the same cells directly removed from the patient cancer. How can I compare the similarities of the two?

Another example: I have cell A transcriptome, and I think that cell A contaminated cell B. Can I compare the cell A transcriptome and cell B transcriptome to infer contamination (or similarity)?

I know there are tools to find immune cell infiltration in tissues based on transcriptome analysis, but how can I compare the similarities between two or more transcriptomics and not a database?

I have found a little about spearman's correlation (which I think would be very weak in whole-transcriptome comparison), but I also found another tool named TROM (https://github.com/Vivianstats/TROM) which does that.

Also I found about funrich, where I can compare my transcriptome to other databases based on more features (proteins etc).

Is there any other method to do this?

[Pratik]

Let's say I have a transcriptome of one expanded cancerous cell, and also I have the transcriptome of the same cells directly removed from the patient cancer. How can I compare the similarities of the two?

You could plug in all of your samples' top variable genes and see where the differences and similarities are through a heatmap.

The tutorial is here:

https://mkempenaar.github.io/gene_expression_analysis/chapter-5.html#clustering

So for example cell A in column 1, cell B in column 2, and suspected contaminated cell (A+B) in column 3., genes on row names. and then see similarities and differences.

Maybe someone could chime in about the statistics, to prove that the observations on the heatmap are significant (I think some of the tools used to do the differential gene expression analysis generate p values for you all together, then it's just a matter of plotting - I am planning to do something similar soon, I can share a response when I do if no one else responds)

Also I found about funrich, where I can compare my transcriptome to other databases based on more features (proteins etc).

Do you mean something like this: https://string-db.org/

You could probably extract your top differentially expressed genes and paste them in there to see protein relationships. Funrich looks cleaner though, imo.

Just some ideas.