Improved documentation

[jguhlin]

Hello, is it possible to get some additional documentation? For example, I do not know what he, he-ai, brent models are. Same for MR, SR, BF. Also, are the defaults always the first possible argument? So MLM is the default model?

Cheers,
--Joseph

[YinLiLin]

Hi Joseph,

Thank you for the question.
Sorry about that, I will add more descriptive details for some major parameters of KAML on the README page, before that, I would like to give a rough explanation about the parameters you mentioned above. "he", "he-ai" and "brent" are the options of algorithms for variance components estimation, "he" represents HE regression, it's super fast, but less of accuracy and stability, "he-ai" represents AIREML with prior values provided by HE regression, and 'brent' is a method of interval estimation of heritability, is more efficient than AIREML by aid of Eigen decomposition technology on genomic relationship matrix. We recommend using "brent", as it has the most stable performance from our large data testing by far. "MR", "SR" and "BF" are the options of how to choose the posterior QTNs, "MR" represents multiple regression, "SR" represents step regression, and "BF" represents using Bonferroni threshold directly.
You are right, if no parameters are specified, KAML will use the first argument for each parameters, "MLM" is the default model, you can find the final used option of each parameters from the printed LOG information.

Best,
Lilin

[jguhlin]

Thanks for the explanations here. It's much appreciated. I'm working on a smaller dataset so the options are very important to adjust.

Cheers,
--Joseph