Individual based distances

NESCent / popgenInfo

Vignettes for Population Genetics in R

http://popgen.nescent.org

MIT License

21 stars 50 forks source link

Individual based distances #46

Open zkamvar opened 9 years ago

zkamvar commented 9 years ago

By marker type:

[ ] SNP
[ ] Sequence
[ ] Microsatellite

Please refer to the issue number when making a pull request.

Questions to address:

What individual based distances exist?
What are their assumptions?
What data types to they require?
How can you use these distances for analyses?
- Clustering
- Mantel Tests
- Dendrograms
- etc.

smanel commented 9 years ago

Examples of individual distances (not exhaustive): Bray curtis distance (no assumptions) and AMOVA distance (gstudio) PCoA or PCA coordinates (adegenet)

I use them mainly for Mantel tests. I think that your are correct in the other applications.

smanel commented 9 years ago

I am working on the vignette individual genetic distance for SNP.
You want the questions be addressed within the vignette or in a more general way?

Just another remark: in the user section, I am wondering if it would not be more easy to replace workflow by the topic/biolgical questions that we have already. Workflow was a basic idea, but they are a little complicate to create. They will be within vignette.

zkamvar commented 9 years ago

Hi Stéphanie,

The questions I wrote here were initial questions that I thought might go well under this topic. I think they should be addressed in any way that feels proper.

Additionally: Correct me if I'm wrong, but the Bray Curtis distance for genetic data is the fraction of dissimilar alleles between two samples? If so, the poppr function diss.dist calculates this quantity.

Regarding the user section, I will create a new issue on that where we can discuss.

smanel commented 9 years ago

For Bray Curtis distance: yes it is. It was used in ecology to quantify differences in species composition.