BMI as a continuous variable is my outcome, which type data i should put to calculate

Question

[jiefei]

BMI as a continuous variable is my outcome, if I want to calculate significant genes, for example, formula like this, BMI ~genes + age + sex, I want to use linear regression model and negative binomial model, when I put genes data, what should I use? CPM?TPM? Or just normalization count?

[ATpoint]

What is the setup here? You have RNA-seq and want to find differences between BMI groups?

[jiefei]

yes, you are right.