[adegenet-forum] spca with type 5

[Thibaut Jombart]

Hi,

For your first question, 'dmax makes no sense - why', this is a tough one,
as one of the two possibles answers is me confessing I'm an idiot. Which
happens, every now and then. This said, the original reason for this is
that by choosing different neighboring windows using dmin and dmax one can
run a spca using the same principle as in correlograms, i.e. varying the
notion of neighourhood at different spatial scales.

As for the rest:
- "sites" = "locations" = spatial coordinates of the statistical units
(individuals, or populations)
- distances are defined as Euclidean distances based on the above
coordinates

Cheers
Thibaut


--
Dr Thibaut Jombart
Lecturer, Department of Infectious Disease Epidemiology, Imperial College
London
Head of RECON: repidemicsconsortium.org
WHO Consultant - outbreak analysis
https://thibautjombart.netlify.com
Twitter: @TeebzR
+44(0)20 7594 3658

On 28 May 2018 at 18:57, Newton, Erica (MNRF) <Erica.Newton at ontario.ca>
wrote:

> Hello;
>
>
>
> I am using the spca command with the “neighbourhood by distance” graph
> (type = 5). In the call to spca, d1 and d2 are referred to as the “minimum
> [or maximum] distance between any two neighbours”, respectively. After
> digging some more I found that in the chooseCN function d1 and d2 are
> defined as:
>
>
>
> d1 = the minimum distance between any two neighbours.
>
> d2 = the maximum distance between any two neighbours. Can also be a
> character: "dmin" for the minimum distance so that each site has at least
> one connection, or "dmax" to have all sites connected (despite the later
> has no sense).
>
>
>
> I am looking for clarification on these arguments. First, “dmax” is an
> option but makes no sense – why?
>
> Second – what are the units of d1 and d2, and do they change based on
> whether your x and y values are, for example, latitudes and longitudes or
> utm coordinates (metres)? What defines a “neighbour”?
>
>
>
> I’m asking because the chooseCN function help page seems to refer to
> “sites” as neighbours – e.g., neighbouring populations or sampling sites.
> When I looked at the spca tutorial, though, it seems like d2 is based on
> the distance between individuals (page 37, “For these data, spatial
> connectivity is best defined as the overlap between home ranges of
> individuals, modelled as disks with a radius of 1150m. chooseCN is used to
> create the corresponding connection network”. d2 is defined as 2300 in the
> call to “chooseCN” in that tutorial.
>
>
>
> I’m asking because I want to ensure that I am choosing the “correct”
> distance for d2. I am working with turtles, who do not have defined
> territories. Some turtles were collected only meters from other turtles,
> whereas others were collected hundreds of kilometers away across a large
> area. We suspect that there are 2-4 “populations” of these turtles from the
> STRUCTURE results.
>
>
>
> Thanks for any clarification you can provide on how to specify d1 and d2
> correctly for my study species. I am new to genetic analyses.
>
>
>
> Best,
>
> Erica
>
>
>
> _______________________________________________
> adegenet-forum mailing list
> adegenet-forum at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/
> listinfo/adegenet-forum
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180530/8fe4dc8e/attachment.html>