How to analyze this data?

[Santhosh Nathan]

I have five unique categorical variables which I would like to present on the one axis (e.g 5 memory loss domains and severity ranging from 0 to 3, continuous, within each domain). There is one variable for the other axis and it has discrete values of either 0, 1 or 2 (a genotype status/copies).

How do I report if genotype status is correlated with each memory loss domain and severity? E.g Does having more genotype copies correlate with higher severity within each domain?

What is the best statistical approach to study this?

[Dave Airey]

Usually predictor SNPs are coded continuously, at least in GWAS contexts, e.g., 0, 1, 2, and for continuous outcomes a regression is run, and the test is for the slope effect per minor allele.

Depending on the mode of action you may also code the genetic data as categorical by genotype, as dominant or recessive, or as a co-dominant with additive and dominance parameters.

A possible issue with your memory loss domains for regression is the measure is continuous but bounded, so regression may predict -1 or +4 for example. You may have to transform or choose an estimator other than regression. I don't know the distributions of the memory loss measures.