EMMAX case-control analysis

[Nick]

Hi,

I have some quick questions regarding EMMAX-based analyses in EPACTS:

1. Which type of kinship matrix is produced by epacts make-kin? Would it be IBS, Balding-Nichols or some other type?
2. Can the single marker EMMAX test (q.emmax) be used for binary traits (case-control analysis)? According to the documentation of EMMAX (http://genome.sph.umich.edu/wiki/EMMAX) the p-values should be reliable but not the effect sizes... does this hold true for the implementation in EPACTS?
3. Similarly, are the p-values produced by the group-based "SKAT using EMMAX" test (mmskat) reliable for a binary trait?

Apologies if I have missed this information in the documentation.

Thanks in advance.

Nick

[Hyun Min Kang]

Hi Nick,

1. It's using centered and normalized (Balding-Nicols) matrix.

2.3 That's correct

--
You received this message because you are subscribed to the Google Groups "EPACTS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to epacts+un...@googlegroups.com.
To post to this group, send email to epa...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/epacts/abcd0ef2-1a48-491f-a326-974b69e59882%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Nick]

Thanks for your quick response.

I assume that by 'correct' you mean the p-values should be OK in these scenarios :-)

Best wishes,

Nick

[Hyun Min Kang]

That's correct :)

--
You received this message because you are subscribed to the Google Groups "EPACTS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to epacts+un...@googlegroups.com.
To post to this group, send email to epa...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/epacts/d0b625e1-dee0-4100-8581-1562c3501cef%40googlegroups.com.