How to use this package?

[biostat0903]

Hi Xiang,
I have read your AOAS paper, and I think ldshrink package is very helpful to me. But when I download the package, it is hard for me to find the main function to calculate the shinkage LD structure. Would you like to give me a toy example to let me use the package?
Many thanks for your kindly help.

                                              Sheng

[xiangzhu]

Dear Sheng @biostat0903,

Many thanks for your interest in ldshrink!

This package is still work in progress, and so the documentations and examples are still lacking. (I originally planned to finish up the package by Thanksgiving, but unfortunately was caught up in other more pressing research assignments.)

I will send you a complete example as soon as the package is (almost) ready.

In the meanwhile you can take a look at this function: https://github.com/stephenslab/rss/blob/master/misc/get_corr.m. This function illustrates how to "calculate the shinkage LD structure", and has been extensively tested in simulations and real data analyses.

[xiangzhu]

This issue can be summarized as:

• add documentations for the functions (see Issue Revise object documentation based on Wickham book #25 );
• add vignettes to illustrate the functions (see Issue Revise vignettes based on Wickham book #26 ).

[biostat0903]

Hi Xiang,

Many thanks for your kindly help. I will do some modification of the get_corr.m.

                                      Sheng

[biostat0903]

Hi Xiang,
I have another question. Would you like to tell me more details of the mean of parameter "Ne". Is it mean the number of populations in the reference panel? For example, I use the 503 samples from 1000 Genome data, the Ne is set as 1.
Many thanks for your kindly help of my project.
Sheng

[xiangzhu]

@biostat0903 Thanks for your question.

As stated in the code (https://github.com/stephenslab/rss/blob/master/misc/get_corr.m), Ne denotes "the effective population size (diploid)".

It is not "the number of populations in the reference panel".

Please also look at this tutorial: https://stephenslab.github.io/rss/recombination.html.

[xiangzhu]

Hi Sheng @biostat0903,

I hope all is well.

In case you are interested, I recently wrote a simple R function to implement the shrinkage LD matrix estimator. You can find the R code here: https://github.com/stephenslab/rss/blob/master/misc/get_corr.R.

Please let me know if you have any question about this function.

I will try to squeeze more time this summer to make some non-trivial progress on ldshrink. I will keep you posted.