How to handle missing population data running amova.test

[Nikki Lukasko]

Hello,

I am trying to run the following amova test:

amova.result_CCR <- poppr.amova(vcf_Bcin_genlight, ~CCR)

I have no problems running it on most populations, but when I try to run "CCR", I get this error:

Error in ade4::amova(samples = xtab, distances = xdist, structures = xstruct) :
  Non convenient samples

In addition: Warning message:
In poppr.amova(vcf_Bcin_genlight, ~CCR) :
  Missing data are not filtered from genlight data.

My column titled CCR contains values 0-7, but has missing data because not all isolates were phenotyped. Is there a good way to subset the data or ignore missing data? I tried looking through previous conversations but nothing I have tried has worked.

Thanks in advance,

Nikki

[Zhian Kamvar]

CCR is the population factor for the model. AMOVA explicitly needs a population factor in order to give you information about the partitioning of the variance. The only way you can subset this is if you exclude the samples with missing data.

You might also be able to use DAPC to impute the population for these samples by assigning the missing values a new value and seeing where they fall based on the analysis. 

I hope that helps. 

--
You received this message because you are subscribed to the Google Groups "poppr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to poppr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/poppr/b72ab1d4-3995-49f2-8762-156de1f4ac6bn%40googlegroups.com.