From thibautjombart at gmail.com  Thu Jan  4 12:49:37 2018
From: thibautjombart at gmail.com (Thibaut Jombart)
Date: Thu, 4 Jan 2018 11:49:37 +0000
Subject: [adegenet-forum] POV in DAPC
In-Reply-To: <AF5DED5D-FFBF-4218-82ED-5C400F61A105@gmail.com>
References: <AF5DED5D-FFBF-4218-82ED-5C400F61A105@gmail.com>
Message-ID: <CANPRA+pZvzrbAFw24KmPOWw=cgrv+M933qurLELROAfepQQXww@mail.gmail.com>

Hi there,

These values are in general percentage of inertia, which in the case of the
PCA is a variance, and in the case of the DAPC is the ratio variance
between / total variance (ie F statistic).

Best
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 13 December 2017 at 13:08, Andrew Veale <andrew.j.veale at gmail.com> wrote:

> Hey!
>
> I have been asked to add the percentage of variance explained by each axis
> in the DAPC scatter plot.  I?m not sure how best to do this, and I think it
> isn?t the most meaningful thing to do as it isn?t a straight PCA.  The POV
> of a PCA is fine, and I can see how to get the Eigen-values for each PC
> retained - just not the then combined ones for the DAPC.
>
> Any thoughts on this?
>
> Thanks!
> _______________________________________________
> 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/20180104/e769141b/attachment.html>

From michaeldonaldson at trentu.ca  Fri Jan 19 16:46:52 2018
From: michaeldonaldson at trentu.ca (Mike Donaldson)
Date: Fri, 19 Jan 2018 10:46:52 -0500
Subject: [adegenet-forum] Do you need to replace missing data before
	performing a glPca?
Message-ID: <CAL-1RbesP0nSvTPYHVQtOMQqGQ-S+gqEzTh7F69rXdEOm7kfRQ@mail.gmail.com>

Hello,

I imported a vcf file and converted it to a genlight object using vcfR as
follows:

gl.x <- vcfR2genlight(file.vcf)
ploidy(gl.x) <- 2
pop <- read.csv("popfile.csv", sep=",", header=TRUE)
pop(gl.x) <- pop$Pop

The dataset has ~10% missing data. When using the glPca function, do you
need to first convert the missing data to the mean and transform it based
on frequency using the tab function (tab(gl.x, NA.method="mean", freq=TRUE)),
or does that happen "behind the scenes"?

I ask because the glPca will proceed without that conversion, while a
dudi.pca will not.

Thank you for your time,
Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180119/075c43c3/attachment.html>

From roman.lustrik at biolitika.si  Fri Jan 19 17:55:51 2018
From: roman.lustrik at biolitika.si (Roman =?utf-8?Q?Lu=C5=A1trik?=)
Date: Fri, 19 Jan 2018 17:55:51 +0100 (CET)
Subject: [adegenet-forum] Do you need to replace missing data
	before	performing a glPca?
In-Reply-To: <CAL-1RbesP0nSvTPYHVQtOMQqGQ-S+gqEzTh7F69rXdEOm7kfRQ@mail.gmail.com>
References: <CAL-1RbesP0nSvTPYHVQtOMQqGQ-S+gqEzTh7F69rXdEOm7kfRQ@mail.gmail.com>
Message-ID: <1874430037.1113289.1516380951441.JavaMail.zimbra@biolitika.si>

Could it be that it goes through because NAs are replaced by 0 ? 

---- 
In god we trust, all others bring data. 


From: "Mike Donaldson" <michaeldonaldson at trentu.ca> 
To: adegenet-forum at lists.r-forge.r-project.org 
Sent: Friday, January 19, 2018 4:46:52 PM 
Subject: [adegenet-forum] Do you need to replace missing data before performing a glPca? 

Hello, 

I imported a vcf file and converted it to a genlight object using vcfR as follows: 

gl.x <- vcfR2genlight(file.vcf) 
ploidy(gl.x) <- 2 
pop <- read.csv("popfile.csv", sep=",", header=TRUE) 
pop(gl.x) <- pop$Pop 

The dataset has ~10% missing data. When using the glPca function, do you need to first convert the missing data to the mean and transform it based on frequency using the tab function ( tab(gl.x, NA.method="mean", freq=TRUE) ), or does that happen "behind the scenes"? 

I ask because the glPca will proceed without that conversion, while a dudi.pca will not. 

Thank you for your time, 
Mike 

_______________________________________________ 
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/20180119/42f5536e/attachment.html>

From mike.donaldson at gmail.com  Fri Jan 19 16:42:21 2018
From: mike.donaldson at gmail.com (Mike Donaldson)
Date: Fri, 19 Jan 2018 10:42:21 -0500
Subject: [adegenet-forum] Do you need to replace missing data before
	performing a glPca?
Message-ID: <CAL-1RbeqGFiMU_Z3hoHBs9OMLyMyKY9Ay4A+XLWNbJkTa47oWg@mail.gmail.com>

Hello,

I imported a vcf file and converted it to a genlight object using vcfR as
follows:

gl.x <- vcfR2genlight(file.vcf)
ploidy(gl.x) <- 2
pop <- read.csv("popfile.csv", sep=",", header=TRUE)
pop(gl.x) <- pop$Pop

The dataset has ~10% missing data. When using the glPca function, do you
need to first convert the missing data to the mean and transform it based
on frequency using the tab function (tab(gl.x, NA.method="mean",
freq=TRUE)), or does that happen "behind the scenes"?

I ask because the glPca will proceed without that conversion, while a
dudi.pca will not.

Thank you for your time,
Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180119/9e6a71cb/attachment.html>

From sheamaddocklambert at gmail.com  Sun Jan 28 23:11:54 2018
From: sheamaddocklambert at gmail.com (Shea Lambert)
Date: Sun, 28 Jan 2018 15:11:54 -0700
Subject: [adegenet-forum] snapclust: NaN for proba
Message-ID: <CAKK1bNpdDAJ_q=edGgRqJj2jjbW=fc+Jetakf3RgRxFxzA9Hsg@mail.gmail.com>

Hello,

I've been trying snapclust, and it runs/converges successfully, but all my
entries for $proba are "NaN". I've tried removing all missing data, but
still get "NaN" for everything. Any advice much appreciated.

Shea
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180128/63893919/attachment.html>

From thibautjombart at gmail.com  Tue Jan 30 13:45:31 2018
From: thibautjombart at gmail.com (Thibaut Jombart)
Date: Tue, 30 Jan 2018 12:45:31 +0000
Subject: [adegenet-forum] snapclust: NaN for proba
In-Reply-To: <CAKK1bNpdDAJ_q=edGgRqJj2jjbW=fc+Jetakf3RgRxFxzA9Hsg@mail.gmail.com>
References: <CAKK1bNpdDAJ_q=edGgRqJj2jjbW=fc+Jetakf3RgRxFxzA9Hsg@mail.gmail.com>
Message-ID: <CANPRA+osgwUs30y=r0qUvz3mSOyht_ufJoFc2Mhn+2qB+-=DGQ@mail.gmail.com>

Hi Shea

my guess would also be missing data. Could you post that as an issue on
github with a reproducible example?

https://github.com/thibautjombart/adegenet/issues

I'll look into it asap.

Best
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 January 2018 at 22:11, Shea Lambert <sheamaddocklambert at gmail.com>
wrote:

> Hello,
>
> I've been trying snapclust, and it runs/converges successfully, but all my
> entries for $proba are "NaN". I've tried removing all missing data, but
> still get "NaN" for everything. Any advice much appreciated.
>
> Shea
>
> _______________________________________________
> 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/20180130/023a8e6c/attachment.html>

From benjamin.dauphin at wsl.ch  Wed Jan 31 10:17:51 2018
From: benjamin.dauphin at wsl.ch (Benjamin Dauphin)
Date: Wed, 31 Jan 2018 10:17:51 +0100
Subject: [adegenet-forum] Kmeans and DAPC on poolSeq data
Message-ID: <22A6ABF6-1D2B-4DB6-9D52-5899300649A8@wsl.ch>

Dear all,

I am newly working on pool sequencing data and I simply wonder if I can use kmeans (find.cluster) and DAPC to investigate population structure from poolseq data (allele frequencies)? How find.clusters can deal with allele frequencies? 

Dataset: 7 pools and 100?000 SNPs

Any comment or help would be much appreciated.
Best regards
Ben



From nlv209 at hotmail.com  Tue Jan 30 19:08:20 2018
From: nlv209 at hotmail.com (Nikki Vollmer)
Date: Tue, 30 Jan 2018 18:08:20 +0000
Subject: [adegenet-forum] How to interpret Density Plot for K=2
Message-ID: <BN6PR06MB3266D6C050395557B388697C80E40@BN6PR06MB3266.namprd06.prod.outlook.com>

Hi,


I am trying to analyze ~200 RADseq loci for ~200 individuals.  STRUCTURE results suggest the best number of populations given the data is 2.  Pairwise Fst values are quite low for my taxa (<0.003) with pvalue 0.01802.  I was trying to do a DAPC on this same data to compare results. DAPC similarly suggested the best # of clusters is 2 and I was able to plot a 1-dimensional density plot for the one DF I kept (attached).  However, I am not sure how to interpret the plot.  Is it correct to say that because the two peaks do not overlap that suggests the 2 clusters are quite differentiated from one another (similar to two clusters on a scatter plot being in opposite quadrants)?  (...or is that logic flawed?)


I am trying to figure out if these 2 groups are very genetically differentiated or not, and I am not clear what the density plot is supporting/suggesting.


I very much appreciate any guidance on this matter!


Thank you,

Nikki

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180130/da47f169/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Rplot.pdf
Type: application/pdf
Size: 1252439 bytes
Desc: Rplot.pdf
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20180130/da47f169/attachment-0001.pdf>

From danielledanielle89 at gmail.com  Wed Jan 31 01:18:46 2018
From: danielledanielle89 at gmail.com (Danielle Louise)
Date: Wed, 31 Jan 2018 10:18:46 +1000
Subject: [adegenet-forum] snapclust
Message-ID: <BDCFDBC3-73B7-4AB3-8E74-1576D52D4F99@gmail.com>

Hello. I am looking at implementing your snapclust function, and I am reading through your recent paper. 

 I have a few questions regarding incorporating empirical data. I have simulated data sets with parental and F1 F2 and BC and I am wondering how to incorporate the empirical data - do I add it in to the simulated data and measure the accuracy of the assignment to classes to then determine the reliability of detection of hybrids in the empirical data? The tutorial gives a good outline of using the simulated data, but I think I am missing something when it comes to checking the empirical data, so I am asking for some really practical advice about how to incorporate the empirical data ?  Also should we bootstrap the final probabilities to clarify the results?

Thanks
Dan