Using DADI for recent population splits

[Mark Derbyshire]

Dear DADI users,

I apologise for the extremely broad nature of this question but I am another population genetics newbie. I am perfectly fine with coding and file formats etc. but hopeless at mathematics. I really like the idea behind the software and the potential flexibility in implementation.

My question is:

I have samples of two populations of a single haploid species of fungus, inferred from there being two broad clades in a phylogenetic tree. Each clade is specific to one side of the world - one in Australia and one spanning Europe and North America.

The fungus was likely introduced to Australia in the last century, possible even the last 50 years, although this is not a certainty. It undergoes a single generational cycle every year, environmental conditions permitting, and outcrosses infrequently. So there is not that much recombination occurring. I have samples of 10 and 12 individuals per population and should be able to infer ancestral alleles from two outgroups.

Given this information, is dadi likely to be useful for assessing the genetic history of the world population in terms of population splits, migration and bottlenecks?

My purposes are twofold - 1) I'd like to gain some information on the demographic history of the fungus, 2) I'd like to simulate data under a similar model to create a null distribution against which to test for selective sweeps.

Any information or advice would be HUGELY appreciated.

Kind regards,

Mark

[Gutenkunst, Ryan N - (rgutenk)]

Hi Mark,

With limited recombination and only 50 generations to work with, I’m not optimistic for the power of dadi in this situation. If the effective population size is low, then drift will have been fast, which will help. But you might be better off with a haplotype-based method, such as IM.

Best,

Ryan

--
You received this message because you are subscribed to the Google Groups "dadi-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dadi-user+...@googlegroups.com.
To post to this group, send email to dadi...@googlegroups.com.
Visit this group at https://groups.google.com/group/dadi-user.
For more options, visit https://groups.google.com/d/optout.

--

Ryan Gutenkunst
Associate Professor of Molecular and Cellular Biology, University of Arizona
phone: (520) 626-0569, office: LSS 325, web: http://gutengroup.mcb.arizona.edu

Latest papers: 
“Selection on network dynamics drives differential rates of protein domain evolution”
PLoS Genetics; http://dx.doi.org/10.1371/journal.pgen.1006132
"Inferring demographic history using two-locus statistics"
Genetics; http://doi.org/10.1534/genetics.117.201251

[Mark Derbyshire]

Hi Ryan,

Excellent, thank you for your prompt reply. 

Kind regards,

Mark

[Gutenkunst, Ryan N - (rgutenk)]

Hi Mark,

It’s not clear to me what the problem is. Try the other optimizers, or other initial parameter settings.

Best,

Ryan

On Aug 16, 2017, at 7:56 PM, Mark Derbyshire <markc...@gmail.com> wrote:

Hi Ryan,

Me again. I have reconsidered my hypothesis, and it may well be the case that there has been a more ancient split between my samples. Long story short, differentiation seems to be present across the native range of the species and not just between Europe and Australia. There are two subpopulations, one of which is present in North Africa/Mediterranean and Australia and the other in Europe and North America.

My problem currently is with optimisation.

I have a split followed by migration model that produces a superficially similar frequency spectrum to the data (see attached). The real data are in the 'datasfs.png' file.

However, when I try to optimise using the input parameters for the attached plot as starting parameters, the optimiser returns an array of the same parameters and nothing else. I am using dadi.Inference.optimize().

Optimisation enters a phase where it keeps testing the same or very similar parameters:

4241    , -4935.2     , array([ 0.801      ,  1          ,  5          ,  0.55       ])

4243    , -4916.17    , array([ 0.8        ,  1          ,  5.001      ,  0.55       ])

4244    , -4932.74    , array([ 0.8        ,  1          ,  5          ,  0.551      ])

4245    , -4916.36    , array([ 0.8        ,  1          ,  5          ,  0.55       ])

4246    , -4916.36    , array([ 0.8        ,  1          ,  5          ,  0.55       ])

4247    , -4935.2     , array([ 0.801      ,  1          ,  5          ,  0.55       ])

4249    , -4916.17    , array([ 0.8        ,  1          ,  5.001      ,  0.55       ])

4250    , -4932.74    , array([ 0.8        ,  1          ,  5          ,  0.551      ])

Then returns:

array([ 0.8 ,  1.  ,  5.  ,  0.55])

Which are the original parameters.

I have tried this with various input parameters and get the same result.

Is there an issue with optimisation that may cause this?

Kind regards,

Mark

<splitmigration08150551010.png><datasfs.png>

[Mark Derbyshire]

Hi Ryan,

Apologies for this post. I thought I deleted it. I found I was setting the upper bounds of the parameter space too low.

Best wishes,
Mark

[Gutenkunst, Ryan N - (rgutenk)]

Hi Mar,

Thanks for posting your solution to the list. That makes it much easier for other users to solve their own problems in the future.

Best,

Ryan

--
You received this message because you are subscribed to the Google Groups "dadi-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dadi-user+...@googlegroups.com.
To post to this group, send email to dadi...@googlegroups.com.
Visit this group at https://groups.google.com/group/dadi-user.
For more options, visit https://groups.google.com/d/optout.

--

Ryan Gutenkunst
Associate Professor and Associate Department Head

Department of Molecular and Cellular Biology, University of Arizona

phone: (520) 626-0569, office: LSS 325, web: http://gutengroup.mcb.arizona.edu

Latest papers: 
“Selection on network dynamics drives differential rates of protein domain evolution”
PLoS Genetics; http://dx.doi.org/10.1371/journal.pgen.1006132
"Inferring demographic history using two-locus statistics"
Genetics; 

http://doi.org/10.1534/genetics.117.201251