Re: Calulating global fst values

[Jose Luis Mijangos]

Hi Ella,

Could you please extend what you mean about global FST values, perhaps you could provide an example or a citation?

P-values are truncated to two decimals. So, a p-value = 0 means p-value < 0.01.

Cheers,

Luis

On Thursday, 18 August 2022 at 10:09:13 pm UTC+10 ella...@gmail.com wrote:

Hi All,

Im wondering if DartR has the ability to calculate global fst values? and if not, are there any suitable packages which utilises genlight data or have easy ways of converting it?

On another note, my pairwise Fst pvalues all equal to 0. What does this mean? Is it a possible error on my behalf? or is it not significant at all?

Thanks for your help,

Ella 

[Ella Dunn]

Hi Luis,

Here are the output Pvalues I received from my Pairwise fst calculation. Is there a way I can increase the decimal places to accurately determine how significant they are from one another?

Thanks,

Ella 

--
You received this message because you are subscribed to the Google Groups "dartR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dartr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dartr/778cc61b-64bb-40c6-8744-249754c75964n%40googlegroups.com.

[Jose Luis Mijangos]

Hi Ella,

The gl.fst.pop function is a wrapper for the function stappFst from Package StAMPP. This function uses Bootstrapping to calculate the p-values. The bootstrapping method resamples the individuals "nboots" times and calculates FST every time for each locus. Then all the calculated FST > 0 are summed and divided by "nboots" to get the p-value. In other words, the proportion of bootstrapped FST values equal to zero or less (no differentiation between populations) is used as the P-value for differentiation.

So, if all the bootstrapped FST values are zero, the p-value will be zero.

You can increase the number of bootstraps ("nboots") to be more confident in the statistical significance of your results.

Cheers,

Luis

[Bernd.Gruber]

Hi,

Not sure if this is what you are looking for.

You can calculate overall Fsts using package hierfstat.

see tutorial here: https://cran.r-project.org/web/packages/hierfstat/vignettes/hierfstat.html

hierfstat uses genind objects so you need to convert you data with the gl2gi function first and then the hierfstat function should work.

If it works can you please let us know.

Cheer, Bernd

—————————————————————————————————————

Bernd Gruber

Professor Spatial and Ecological Modelling

Institute for Applied Ecology 

P 02 62063804  E bernd....@canberra.edu.au

appliedecology.edu.au

University of Canberra, ACT 2617

NOTICE & DISCLAIMER: This email and any files transmitted with it may contain confidential or copyright materials and are for the attention of the addressee only. If you have received this email in error please notify us by email reply and delete it from your system. The University of Canberra accepts no liability for any damage caused by any virus transmitted by this email.Australian Government Higher Education (CRICOS)Registered Provider number: #00212K

—————————————————————————————————————

On 20 Aug 2022, at 10:20, Jose Luis Mijangos <luis.m...@gmail.com> wrote:



--

[mukiibi robert]

Hi Luis,

I am kind of having the same problem. I ran the following command to estimate Fst values >>>fst2=gl.fst.pop(snp_data, nboots = 100, percent = 95, nclusters = 1, verbose = NULL).

All my estimates were returned with Pvalues 0, even those that are very low such as 0.01. I am wondering if there is a way to increase the number of decimal places dartR can return for the pvalues. I have tried higher  nboots but no change observed. I am a little skeptical to refer to an Fst of 0.01 as significant, however one of our reviewers wants us to attach pvalues to our Fst estimates.

Thank you in advance,

Robert

[Jose Luis Mijangos]

Hi Robert,

As mentioned, in the previous conversation, the p-value is 0 when none of the bootstraps results is less than zero, so there is not possible to increase the decimals when the p-value is 0. 

Another approach that might be more useful is to calculate Confidence intervals, as shown below. 

devtools::install_github("green-striped-gecko/dartR@dev")

res <- gl.report.fstat(platypus.gl,
                       nboots = 2000,
                       CI.type = "perc",
                       ncpus = parallel::detectCores()-1 ,
                       )

Cheers,

Luis