LOCPRIOR and POPFLAG: can they be used together?

[Emma Carroll]

Hi,

I am working on a project to identify the source wintering ground for animals from a migratory corridor (mixed sample). The source wintering grounds cluster into two populations using the location prior model in a way that is consistent with the degree of connectivity inferred from the movement of individuals.

The next step is to identify which of the clusters the individuals from the mixed migratory corridor samples come from. As the USEPOPINFO and LOCPRIOR models are not compatible, and the clusters are not resolved using the USEPOPINFO model,
I thought I could:
- set POPFLAG=0 for the migratory corridor samples to ensure the samples do not influence the STRUCTURE analysis (I assume this will work on LOCPRIOR model??)
- run STRUCTURE with migratory corridor samples assigned to one cluster and then assigned to the other cluster
- see if the migratory samples (each) consistently assign to one population in both analyses

We have an internal check on this as we have some direct matches between the migratory corridor and wintering grounds.

Can anyone tell me if my assumption that using the POPFLAG option with the LOCPRIOR model is correct? And if there is a better way to do the analysis I am suggesting?

thanks
Emma

[Vikram Chhatre]

Emma -

As far as I understand, the POPFLAG parameter is specific to USEPOPINFO model and unless sample locations are going to be used for learning purposes by the STRUCTURE algorithm, POPINFO is to be disabled.

Thus, I see no reason for the LOCPRIOR model to *not work* when POPINFO is disabled.  Note that LOCPRIOR model can either use the data from POPDATA column (LOCISPOP=1) or from a separate LOCDATA column.  See manual for details.

Let us know what you find and/or if you have further questions.

Vikram

Emma

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

[Vikram Chhatre]

Edit: I meant to say: I see no reason for the LOCPRIOR model to *not work* when POPFLAG is disabled.  

Sorry about the typo.

V

[Charalambos Neophytou]

Hi Vikram,

now I have a similar problem, but I'm not sure I fully understood the conversation above. My question is: Is there reason for the LOCPRIOR model to *not work* when POPFLAG is *enabled*?

What I'm trying to do is to compare introduced populations (in Europe) with native populations of a forest species (a North American one). So after applying LOCPRIOR within the native populations I found some clusters (and of course some admixed populations) which are meaningful (in agreement with geography, historic migration etc.). When including introduced (European) populations, I thought it could help to set POPFLAG to 1 within those native populations that showed high affinity to one cluster (say q > 0,90).

Keeping LOCPRIOR would still make sense since the genetic differentiation within the species (also among native populations) is very low (Fst = 0,02-0,03...).

Also, each introduced population is more likely to be of one origin (LOCPRIOR again meaningful).

So, can I use LOCPRIOR and USEPOPINFO simultaneously.

Thanks and best regards,

Charalambos

To unsubscribe from this group and stop receiving emails from it, send an email to structure-software+unsub...@googlegroups.com.

[Vikram Chhatre]

Hi Charalambos,

Both LOCPRIOR and USEPOPINFO are models used for using prior sample information in 'learning samples' to assisting the ancestry inference of the remaining samples. That said there are some differences between LOCPRIOR and the two ways USEPOPINFO can be used (all three described in the user manual).

Given the information you provided about your native and introduced samples, the USEPOPINFO model 

"pre-specify the population of origin of some individuals to assist ancestry estimation for individuals of unknown origin"

seems to be appropriate for the analysis of your data.  Here you would set POPFLAG=1 in the parameters, create a new locdata column where you would assign 1 to native samples and 0 to introduced samples.  

That way, you are telling structure that you already know what clusters the native samples belong to.  This information will be used in the inference about the introduced samples whose population location information and grouping will be completely ignored by the model.

I hope that clarifies your doubts.  We would love to see your results/bar plots once you are through.

V 

To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.

To post to this group, send email to structure...@googlegroups.com.
Visit this group at http://groups.google.com/group/structure-software.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "structure-software" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.

To post to this group, send email to structure...@googlegroups.com.

Visit this group at https://groups.google.com/group/structure-software.
For more options, visit https://groups.google.com/d/optout.