Training set

[austnle]

Hi Dayoon!

Hope you're doing well and thanks for sharing this package. I understand that for hd_calc, the training dataset is NHANES III with individuals 20-30 years of age, a hypothetically young, "healthy" cohort, however for the phenoage_nhanes and kdm_nhanes, I am having a hard time understanding the NHANES III training cohort used for these functions. Are these populations also assumed to be "healthy agers"?

Also, in your code to create a table:

table3 = table_ses(data,agevar,exposure = c("edu","annual_income","poverty_ratio"), label)

#pull table
table3$table

you mentioned that "Models included covariates for chronological age and sex. The original KDM Biological Age algorithm (left-most column) was projected onto data from NHANES 2007-2010 only because other NHANES IV waves did not include spirometry measurements. The original Levine Phenotypic Age algorithm (second column from left) was projected onto data from NHANES 1999-2010 and 2015-2018 only because the intervening waves did not include CRP measurements."

How did you specify model covariates and the projection data into the table_ses function?

[yingjielixy]

Hello, this package is very convenient for calculating epigenetic age in NHANES. Many thanks to the author for their efforts. I have a similar question: I noticed that the KDM age I calculated tends to be significantly lower, about ten years younger compared to PhenoAge. Is this normal? Additionally, can this package be applied to the newly released NHANES data cycles?