imperfect microsat coding

[Stu Willis]

Hello all,

I'm sorry this is a naive question: with the implementation of the
fragment length coding scheme for microsatellites, are we still tied
to rounding imperfect repeat alleles (i.e. 1bp difference) up or down
to fit to stepwise or Brownian models?

I have 2 out of 12 loci which show several alleles with 1bp
differences (otherwise di-nucleotide), although these alleles
fortunately make up the minority for each locus (24 or 50 observations
out of 272). More than bias introduced by rounding, I just hate
throwing away information!

Anyone have advice on how best to deal with this? Thanks!

~Stu

[Peter Beerli]

Stu,
the adjustment of the fragment length is handled independently of the mutation model therefore
you will need to live with the rounding up or down [I may include the correct fractional repeat number into the brownian motion model as that should be easy, but it is not done yet], I bet that this will not make any difference in your answer
if you run this a few time to check 9each time the rounding up or down will be with p=0.5 if it is a dinuc, and for the tetranuc
it depends on how many sites one is off of the perfect repeat, p for rounding down +1->0.75, +2->0.5, +3->0.25 etc

Peter

> --
> You received this message because you are subscribed to the Google Groups "migrate-support" group.
> To post to this group, send email to migrate...@googlegroups.com.
> To unsubscribe from this group, send email to migrate-suppo...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/migrate-support?hl=en.
>

[Stu Willis]

Thanks Peter. So the program will now round automatically? Or is that
a recommendation for how to implement the rounding? Thanks for the
clarification.

~Stu

[Peter Beerli]

Stu
if you add the line below to the data the msat fragment length will be automatically translated into repeats according to the list (the example shows two dinuc and 1 trinuc).

#@M 2 2 3

rounding will be automatic.

Peter