Is the Kruskal-Wallis test appropriate for genomic distances?

Question

[aaronlogsdon2001]

Hi all. I have an independent variable and a categorical variable:

• Dependent variable = genomic distance between retroviral integration site and histone mark H3K36me3
• Independent variable = dataset type (HIV in vivo, HIV in vitro, HTLV-1 in vivo, HTLV-1 in vitro, random)

The genomic distances are in kb (to three decimal places) after I divided the base count by 1000.

Does this mean my dependent variable is continuous, and does this mean I am not violating any of the rules for a Kruskal-Wallis test (and Dunn test post-hoc follow-up)?

[ATpoint]

I never used it myself so far, but I would proably look into https://bioconductor.org/packages/devel/bioc/vignettes/nullranges/inst/doc/nullranges.html to get your stats and choice of background regions right.