how to create a sequence input file for migrate-n?

[jiehan]

Hi,

I have tried to create a sequence input file for migrate-n.exe using "CREATE", but it failed to read FASTAT or ARLEQUIN file, then I did it by hand following MIGRATE manual, please see the attached file, but migrate-n failed to read it. 

Could anybody help me?

Cheers,

Hannah

[andersgs]

Hi Hannah. By default, you need exactly 10 characters separating the beginning of the line and the sequence data. See the attached. Good luck. Anders.

[jiehan]

Hi Anders,

Thank you so much for your kind help! The program is running.

I don't have any experience on sequence data using migrate-n. I acturally have 8 pop and 332 sequences, for the search strategy, I'm using Bayesian
inference, and I set the number of recorded steps in chain to 30000, and burn in is 10000. Do you have any suggestions please?

Cheers,

Hannah

--
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 http://groups.google.com/group/migrate-support.
For more options, visit https://groups.google.com/d/optout.

[andersgs]

Hi Hannah.

There are plenty of discussions in the forum that can help you there. I suspect those runs will be too short. I also think it would be unrealistic to expect Migrate-N to estimate 64 parameters with a single locus. You should try to group populations or reduce the number of migration parameters.

I know nothing about your study system. But, by grouping populations, I mean, that if you have sets of populations that are separated by a river, you might want to group the populations into individuals from either side of the river, and estimate 4 parameters, population size for each side of the river, plus migration in each direction.

As I said, there are plenty of discussions threads in the forum all ready to help you there.

Best of luck.

Anders.

[韩洁]

Hi Anders，

Thanks again! Our study is on a marine clam，and there is no obvious barrier among/between these populations，and currently we onIy have one mtDNA Iocus data. Anyway，I will consult the forum.

Best regards，

Hannah

[Cheng-Min SHI]

Hi Hannah,
    Try PGDSpider (http://cmpg.unibe.ch/software/PGDSpider/) to format your infile for Migrate. The common format mistakes I met when prepare infile manually is  (1) miscount number of sequences for populations, (2) using space instead of tab to as seperator, (3) including special characters, e.g. returns, extra space at the end of sequences. Anyway, I am not absolutely sure yet, whether these mistakes could cause failure to read the infile by Migrate.
    I wonder if it feasible to define a prior the number of population groups using other method, e.g. SAMOVA, so that to reduce number of parameters in your Migrate models as Anders pointed out.
Best

Cheng-Min

> -----原始邮件-----
> 发件人: "韩洁" <jie...@bnu.edu.cn>
> 发送时间: 2015年5月27日 星期三
> 收件人: andersgs <ande...@gmail.com>
> 抄送: migrate-support <migrate...@googlegroups.com>
> 主题: Re:[SPAM] Re: [SPAM] [migrate-support] Re: how to create a sequence input file for migrate-n?
> 
> Hi Anders，

--
Group of Molecular Ecology and Evolution,
Institute of Zoology, Chinese Academy of Sciences,

1 Beichen West Road, Chaoyang District
Beijing 100101, China
Tel: +86 10 64807203
Email: sh...@ioz.ac.cn

[Peter Beerli]

Cheng-Min and Hannah,

concerning errors with hand-created migrate datasets:

(1) “miscount number of sequences and populations" [yes this is very common]

(2) “using space instead of tab a separator” NO! Migrate originally used ONLY spaces, tabs are not officially supported but work most of the time (I added this ’support’ after receiving too many requests for help with infiles that did not work).

(3) returns and extra spaces should not really matter, but you are correct that there are situations where an extra space at the end of the line can lead to problems.

[4] the most common error is to not use 10 characters for the individual name and not knowing how long the sequence is, or (for msat) forgetting that a separator must be specified.

Peter

[韩洁]

Hi Cheng-min，

Thank you for kindly sharing your experience on Migrate. So I know that PGDSpider can be used to create an infile for Migrate. And l shall try SAMOVA to detect potential group and reduce the number of parameters.

Cheers，

Hannah