heatmap comparisons

[papori]

Hi all,
I have 2 heatmaps/ matrices of 150K cells in each.
The size of both matrices equal (rows and columns).
The info is:
patients X genes

Each patient has genes score (this score is the output of a test that we are doing, and the scores are from 1-1000).

I am looking for a way to find the most significnt different/similar columns in both.

I am working on R, but i dont care to do this somwhere else..

I found this to compare 2 heatmaps:

http://cran.r-project.org/web/packages/pairheatmap/pairheatmap.pdf

But honestly, i don't understand how to work with this..

Any ideas?

Thanks,
Pap

[papori]

Maybe the previous post was not clear.
I am jist looking for any ideas on how to compare 2 heatmaps?

Thanks,
Pap

[rhinoceros]

Well, this could be relevant: pairheatmap: comparing expression profiles of gene groups in heatmaps.

[papori]

Well, this could be relevant: pairheatmap: comparing expression profiles of gene groups in heatmaps.

Thanks rhinoceros.
This is the link that i shared also.
Did you try it?

My data is not expression.. its DNA

[rhinoceros]

Hey,

I didn't try it. I'm not sure I understand why you want to compare heatmaps anyway. Wouldn't it be easier to do some plain statistical analysis with the numbers alone?

[RogBasus]

Can you be more clear as to how the data is formatted (perhaps show an example) and what it is you are specifically trying to determine?

[tomc]

Maybe the previous post was not clear.
I am jist looking for any ideas on how to compare 2 heatmaps?

Thanks,
Pap

how about just start very simple and subtract one from the other.

maybe take the absoute value of the resulting ...heatmap

then look at the rows or columns and sum what you find in each
from some point of view the ones with the smallest sum will be more similar
the ones with the larger sums more different.

then maybe think about how unfair it is for a row or column to have a few large values which ruin it for everyone else and maybe find an average for each and compare those.

and after a while when something about how a plain average might not be telling the whole story starts to bug you, study statistics for a bit.