Analyse a virus

[robertkraus]

Hi!

Does anyone here know of a good example or tutorial or whatever sort of advice when using migrate-n for virus sequences? I would like to learn about strategies to make the program fit for this type of data. It will be a (+)ssRNA virus. The sequences are known to be under purifiying selection. Isolates span several decades and originat from many places.

How do I deal with dated samples? What are reasonable priors? What about mutation models? What to do with sequences which are under variying selection pressures depending on the sites?

I don't expect straight answers to any of these questions. It be cool to read a document as an entry to this sort of thinking.

Cheers,
robert

[Pablo G Goikoetxea]

Hi Robert

As far as I know, migration model assumes neutral markers. It seems you should not use sequences known to be under selection (purifiying or whatever type)

Pablo

--
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.

--

-------
Pablo G Goikoetxea
Don Vela 54, 2º izda
01009 Vitoria-Gasteiz (SPAIN)

[Peter Beerli]

Robert,

Does anyone here know of a good example or tutorial or whatever sort of advice when using migrate-n for virus sequences? I would like to learn about strategies to make the program fit for this type of data. It will be a (+)ssRNA virus. The sequences are known to be under purifiying selection. Isolates span several decades and originat from many places.

No good tutorial yet (but I agree with you that I should have on the website). With strong selection I guess all samples are of the selected type, and thus are neutral among themselves, so no problem; problem may appear if there is adaptation to different hosts etc.

How do I deal with dated samples? What are reasonable priors? What about mutation models? What to do with sequences which are under variying selection pressures depending on the sites?

Dated samples will need an additional file, I suggest to use a timescale in generations and then set the generation-time = 1 and also mutation rate as per generation. Mutation model in 3.6.5 is restricted to F84 whereas in 4.0 you can do Tamura-Nei, if you want to use GTR you may want to try BEAST, who certainly as considerable experience working with virus data.

I don't expect straight answers to any of these questions. It be cool to read a document as an entry to this sort of thinking.

Again, have a look at work by Andrew Rambaut, Trevor Bedford, and other BEAST collaborators.

Peter

[Robert Kraus]

Dear Peter,

 

thanks for your response. BEAST is sort of the default for analysing this virus, and I have not worked with BEAST yet. Maybe it is just about time to look into it. But on the other hand I think BEAST does not estimate migration, right? I'd be interested in migration between rather localised occurrences of this vector borne virus. Therefore the search for example papers of migrate-n with virus.

 

robert

 

Gesendet: Donnerstag, 27. November 2014 um 22:41 Uhr
Von: "Peter Beerli" <beerli...@gmail.com>
An: migrate...@googlegroups.com
Cc: "Robert Kraus" <Robe...@gmx.li>
Betreff: Re: [migrate-support] Analyse a virus

[Peter Beerli]

Look at

Bedford Cobey Beerli and Pascual

Peter

(sent from by phone)