Modulation of the average density

Dear flowers,

Could you possibly suggest references to the literature (and ideally
to the corresponding publicly available datasets) that investigate
an influence of the categorical factor (e.g. an exposure to certain
disease) or quantitative variable (e.g. marker of the disease
progression) on the shape of the average multivariate (e.g. 2D)
density of the POPULATION of multiple subjects?

At the moment I know about the following biologically trivial
example:

=====================================================================
The exposure to HIV+ virus (categorical variable) kills CD4+ positive
cells and therefore affects a two-dimensional (CD4, CD8) density of
fluorescence distribution of the T-cells.
=====================================================================

Dr. Mario Roederer generously shared with us datasets illustrating
this effect.

Thank you.

Misha Salganik

Department of Biostatistics
Harvard School of Public Health