Help in Migrate

[Pedro Peres]

Hey, everyone

This is my first time using Migrate and I am having some problems with my dataset.

I have read some previous questions here and tryed to perform new analysis but I keep having the same problem with prior distribution

I am using 10 microsatellite loci and I have 5 populations.

Other types of analyse showed that these populations are not genetically structured, and I want to use migrate to support the conclusion that they are a metapopulation.

1) I choose Bayesian Analysis

2) I used 'allelic data' because my ssr did not fit in the stepwise mutation model

3) I choose full migration

4) I used the M default values, and for theta I changed for min=0, media= 5, max=10, delta=1 (as exponential prior with window distribution) - outfile 3

5) Number of recorded steps in chain = 500000

6) Burn in = 50000

7) Heating = static

8) Multiple markov chains = 4

After that, because I got some problems, I perfomed the analysis again changing just the theta min=0, media=50, max=100, delta=10 (as recomended by the potential problems page in outfile 3 ) - outfile 4

I think that the outfile 3 seems better, but is not correct yet. 
Am I using Migrate correctly? I really don't know how to continue after this last two runnings..

Does anyone have a suggestion?

Thanks in advance!

https://drive.google.com/file/d/0B9fpoIiq5inzRG1FbURGZzhVU2M/view?usp=sharing

https://drive.google.com/file/d/0B9fpoIiq5inzd1c3TmZWVEEtbVk/view?usp=sharing

[Carlo Pacioni]

HI Pedro,

two things you should pay attention to:

1. you are estimating a huge number of parameters. Try to start with a simpler mig model. You could even start simply by setting only one population. If you really want to have more than one pop, would you have a sensible way to divide these populations in a smaller number of group, perhaps 2 to start with? Or a island mig model? That would reduce the number of parameters that migrate is trying to estimate.

2. Multi-modality in your plots suggest that your runs are too short. My opinion is that you don't  really need 500k data points. I think you can drop this (the number of recorded steps) to, say, 20000 and increase the sampling interval to 1000. 

Try these changes and see how you go. You may also want to look at https://github.com/carlopacioni/mtraceR (tutorial attached), which may help to further identify problems in your analysis.

Hope this helps,

carlo

--
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-support+unsubscribe@googlegroups.com.
To post to this group, send email to migrate-support@googlegroups.com.
Visit this group at https://groups.google.com/group/migrate-support.
For more options, visit https://groups.google.com/d/optout.