NULL ALLELE vs missing data

[Vanina Castroagudin]

Hello Zhian and everyone,

How can we code a null SSR allele (no SSR repeat units present in a certain individual) to differentiate it from a zero denoting missing data?

Thanks,

Vanina Castroagudin

[Zhian Kamvar]

Hello,

If I understand this correctly, you want to be able to indicate that an SSR locus is not missing, but it contains zero repeats for the purposes of calculating Bruvo's distance, using the stepwise mutation model and the problem is that  you want to indicate that this allele that this individual/locus with zero repeats is not missing, it just has zero alleles and you want to include it in the calculations (e.g. if you compare it with an allele that contains 2 repeat units, Bruvo's distance will see that there is a 2-step difference).

Of course, we run into the problem where poppr considers "0" to be a missing allele for the purpose of accommodating polyploids with ambiguous dosage information (where Bruvo's distance shines). This convention is taken from the GenAlEx data specifications. 

I think there are two ways you could code this in your input data to recognize the null allele as the presence of an absence:

1. If you are using _repeat lengths_, then you can shift the data by 1 and code all null alleles as 1 and missing data as 0. In this scenario, you would use a repeat length (argument replen) of 1 for all loci in bruvo.dist().

2. If you are using _fragment length_, then you can add in the length of the primer pairs, which will make the null-alleles non-zero, but still effectively zero length compared to the rest. In this scenario, you would need to supply the repeat lengths for each locus in bruvo.dist() (and be sure to make sure to use fix_replen()).

I hope that helps.

Best,

Zhian

--
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/c283a866-de53-4e5c-af43-55360c6c012an%40googlegroups.com.