Details on PGS phenotype transformation

Hello,

I am interested in using the PRS for total triglycerides for a project I am working on. I am however confused on where I can find the details of the training data or the actual methods that were used. I have consulted both the PLOS paper and the Medrxiv preprint to try to track down this information and I have come away empty handed. Specifically, I am interested in learning what GWAS dataset they used for training and whether the observed triglyceride values were log transformed. Since observed triglyceride values are skewed, I have seen many papers either use a log or inverse-normal transform to coerce the observed data to follow the assumptions of linear regression. 

If anyone has any insights into this, it would be extremely helpful.

Thank you