Per chromosome compartment strength with different eigenvectors

[nponts]

Hi!
First, your tool is awesome and I love it :) Thank you soooo much!
But I have a problem and I need help. I have called compartments using gc-oriented eigenvectors, and it happens that for some chromosomes I have to fetch the 2nd eigenvector, otherwise I catch the chromosome arms. From there, I wanted to compute compartment strength and this is where it gets tricky...
1/ how can I call compartment strength genome-wide taking into account that sometimes 1st EV reflects compartmentalisation, other times 2nd EV is the way to go?
2/ can I call compartment strength on a per chromosome basis and using 1st or 2nd EV as needed?
3/ provided I can do 2/, can I compare directly the obtained strength values to state compartmentalisation is different between chromosomes?
Sorry for my silly questions... And thanks again for the provided tool :)

[liz-is]

Hi,

There's some relevant discussion in a previous issue about this. In this situation, I've usually compared the first and second eigenvectors with gene density and histone modification data if available. You can then determine the eigenvector that actually corresponds to the compartmentalisation for each chromosome separately and stitch these together to create a genome-wide eigenvector file to use as input for the compartment strength calculations. Here's some example code for how I've approached this in the past.

[nponts]

Thank you, I'll go from there!
Cheers!