Migrate-N vs. BayesAss

[Juliana Menger]

Dear all,

I'd like to compare contemporary vs historical migration rates between a pair of populations, by using the approach of Chiucchi & Gibbs 2010 Mol. Ecol. (see attached on page 5359, highlighted).

To do so, they take M values from Migrate and divide by a mutation rate of 0.005 (as for microsatellites), to obtain a comparable value of m estimated from Bayesass.

My results from Migrate are:

M2 -> 1 = 13.833 (immigration from pop2 to pop1)

M1 -> 2  = 26.833 (immigration from pop1 to pop2)

and after dividing by 0.005: 

m2 -> 1 = 27666

m1 -> 2  = 53666

However, Bayesass m is scaled from 0-1 and I can't see how I can compare both results.

Any hint is very welcome!

Thanks a lot,

Juliana

 

[Peter Beerli]

Juliana,

migrate measures immigration as the ‘mutation-scaled immigration rate M’ which is the immigration rate m divided by the mutation rate. Thus, a 

M_1->2 = m_1->2 / mu = 13.833 becomes m_1->2 = 0.069165.

This still seems rather high, but you also use a very high mutation rate.

Peter

--
You received this message because you are subscribed to the Google Groups "migrate-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to migrate-suppo...@googlegroups.com.
To post to this group, send email to migrate...@googlegroups.com.
Visit this group at https://groups.google.com/group/migrate-support.
For more options, visit https://groups.google.com/d/optout.
<(Chiucchi & Gibbs 2010) Similarity of contemporary and historical gene flow among highly fragmented populations of an endangered rattlesnake.pdf>

[Juliana Menger]

Thanks Peter!

actually the mutation rate is 0.0005 (I forgot a zero, sorry).

So, following your calculation it would be 0.0069165.

BayesAss estimation is 0.01 (the fraction of individuals in pop 2 that are migrants derived from pop 1).

Are the results equivalent?

Can I say contemporary migration is higher than historical?

Thanks again,

Juliana

[Peter Beerli]

I am sure both programs deliver credibility intervals (migrate does), use them. I guess that they will bracket the 0.01 value and so you would say that the more recent (last 2-3 generations in bayesass) are similar to the long term average (that of course includes the last 3 generations because migrate does not give you the migration rates x generations ago.

Peter