Dmitrii Kriukov

I faced with the same problem. In my case, normalization of durations `T = T/T.max()` led to a convergence of the solution.

Same issue. Is there a way to generalize this trick if I have 8 samples for young and 8 samples for old groups (16 unique groups in total)? I think...

> Thank you for developing this useful package! I am having trouble implementing the DNAmAge function with normalization = TRUE. Running DNAmAge with or without normalization has similar run time...

This is the expected behaviour of clocks bumping with data exhibiting [covariate shift](https://gsarantitis.wordpress.com/2020/04/16/data-shift-in-machine-learning-what-is-it-and-how-to-detect-it/). The clock just do extrapolation of values not seen in the train dataset and returns nonsense result.

Thank you for this suggestion! We will add it in the next package version.

No, data was retrieved from GEO as they are. We found this more logical due to different ways of preprocessing data for clocks construction. Thus, having all data preprocessing methods...