From Landivincenzo at yahoo.it Fri Sep 8 22:58:35 2017 From: Landivincenzo at yahoo.it (Vincenzo Landi, PhD) Date: Fri, 8 Sep 2017 22:58:35 +0200 Subject: [adegenet-forum] PERCENTAGE OF RETAINED VARIANCE PER AXES Message-ID: Hello, how I can calculate the percentage of retained variance for each of two axis dipayed in the plot of DAPC? Thank you Vincenzo -- Vincenzo Landi, PhD Animal Breeding Consulting, S.L. Laboratorio de Gen?tica Molecular Aplicada Empresa de Base Tecnol?gica de la Universidad de C?rdoba Parque Cient?fico Tecnol?gico de C?rdoba c/ Astronoma Cecilia Payne 8?1 Edif. Centauro 14014 C?rdoba (Espa?a) www.abcgenetica.com/ Tf: +34957218708, +34630639439 FAX: +34957218706 . From thibautjombart at gmail.com Mon Sep 11 14:05:46 2017 From: thibautjombart at gmail.com (Thibaut Jombart) Date: Mon, 11 Sep 2017 13:05:46 +0100 Subject: [adegenet-forum] Loci selection for SNP data In-Reply-To: <160565061.1889947.1502109951577.JavaMail.zimbra@insa-rennes.fr> References: <160565061.1889947.1502109951577.JavaMail.zimbra@insa-rennes.fr> Message-ID: Hello, there is no tool specifically dedicated to this, but you could run a PCA, retain axes corresponding to the xxx% of variance, and then keep only the alleles contributing most to these axes. Here's a quick example: > data(sim2pop) > pca1 <- dudi.pca(tab(sim2pop), scannf = FALSE, nf = 3) > contrib <- pca1$c1^2 > which(contrib > 0.01, TRUE) # contrib > 1% row col L02.06 21 1 L04.02 47 1 L05.02 60 1 L05.09 67 1 L06.6 74 1 L07.7 84 1 L08.07 91 1 L09.8 102 1 L09.9 103 1 L10.01 104 1 L10.12 115 1 [...] Best Thibaut -- Dr Thibaut Jombart Lecturer, Department of Infectious Disease Epidemiology, Imperial College London Head of RECON: repidemicsconsortium.org WHO Consultant - outbreak analysis sites.google.com/site/thibautjombart/ Twitter: @TeebzR +44(0)20 7594 3658 On 7 August 2017 at 13:45, Ngoc-Tien Tran wrote: > Hello, > I'm working with SNP data with a large number of loci. > I would like to know if there is a tool in the package "adegenet" allowing > me to select just one smaller set of loci without loosing information. > > Thanks for your help ! > Best wishes, > Ngoc-Tien > > > -- > -------------------------------------------------------------- > TRAN Ngoc-Tien > D?partement G?nie Math?matiques/Analyse de Risque et Optimisation et > Mod?lisation. > INSA-Rennes > Tel : +33.(0)7.51.57.96.53 > @mail : ngoc-tien.tran at insa-rennes.fr > _______________________________________________ > 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: From Ngoc-Tien.Tran at insa-rennes.fr Tue Sep 12 22:19:46 2017 From: Ngoc-Tien.Tran at insa-rennes.fr (Ngoc-Tien Tran) Date: Tue, 12 Sep 2017 22:19:46 +0200 (CEST) Subject: [adegenet-forum] Loci selection for SNP data In-Reply-To: References: <160565061.1889947.1502109951577.JavaMail.zimbra@insa-rennes.fr> Message-ID: <841652196.1979406.1505247586764.JavaMail.zimbra@insa-rennes.fr> Hello , Thank you so much for your suggestion. However, this method only works if the position on the chromosomes are available for each locus. Am I correct ? Best wishes ! Ngoc-Tien ----- Mail original ----- De: "thibautjombart" ?: "Ngoc-Tien Tran" Cc: "adegenet-forum" Envoy?: Lundi 11 Septembre 2017 14:05:46 Objet: Re: [adegenet-forum] Loci selection for SNP data Hello, there is no tool specifically dedicated to this, but you could run a PCA, retain axes corresponding to the xxx% of variance, and then keep only the alleles contributing most to these axes. Here's a quick example: > data(sim2pop) > pca1 <- dudi.pca(tab(sim2pop), scannf = FALSE, nf = 3) > contrib <- pca1$c1^2 > which(contrib > 0.01, TRUE) # contrib > 1% row col L02.06 21 1 L04.02 47 1 L05.02 60 1 L05.09 67 1 L06.6 74 1 L07.7 84 1 L08.07 91 1 L09.8 102 1 L09.9 103 1 L10.01 104 1 L10.12 115 1 [...] Best Thibaut -- Dr Thibaut Jombart Lecturer, Department of Infectious Disease Epidemiology, Imperial College London Head of RECON: repidemicsconsortium.org WHO Consultant - outbreak analysis sites.google.com/site/thibautjombart/ Twitter: @TeebzR +44(0)20 7594 3658 On 7 August 2017 at 13:45, Ngoc-Tien Tran wrote: > Hello, > I'm working with SNP data with a large number of loci. > I would like to know if there is a tool in the package "adegenet" allowing > me to select just one smaller set of loci without loosing information. > > Thanks for your help ! > Best wishes, > Ngoc-Tien > > > -- > -------------------------------------------------------------- > TRAN Ngoc-Tien > D?partement G?nie Math?matiques/Analyse de Risque et Optimisation et > Mod?lisation. > INSA-Rennes > Tel : +33.(0)7.51.57.96.53 > @mail : ngoc-tien.tran at insa-rennes.fr > _______________________________________________ > 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 -- -------------------------------------------------------------- TRAN Ngoc-Tien D?partement G?nie Math?matiques/Analyse de Risque et Optimisation et Mod?lisation. INSA-Rennes Tel : +33.(0)7.51.57.96.53 @mail : ngoc-tien.tran at insa-rennes.fr From thibautjombart at gmail.com Tue Sep 19 16:20:00 2017 From: thibautjombart at gmail.com (Thibaut Jombart) Date: Tue, 19 Sep 2017 16:20:00 +0200 Subject: [adegenet-forum] Loci selection for SNP data In-Reply-To: <841652196.1979406.1505247586764.JavaMail.zimbra@insa-rennes.fr> References: <160565061.1889947.1502109951577.JavaMail.zimbra@insa-rennes.fr> <841652196.1979406.1505247586764.JavaMail.zimbra@insa-rennes.fr> Message-ID: Hi there, no, the position of the loci is not needed. This retains most contributing alleles, no matter what the loci are. Best Thibaut -- Dr Thibaut Jombart Lecturer, Department of Infectious Disease Epidemiology, Imperial College London Head of RECON: repidemicsconsortium.org WHO Consultant - outbreak analysis sites.google.com/site/thibautjombart/ Twitter: @TeebzR +44(0)20 7594 3658 On 12 September 2017 at 22:19, Ngoc-Tien Tran wrote: > Hello , > Thank you so much for your suggestion. > However, this method only works if the position on the chromosomes are > available for each locus. Am I correct ? > > Best wishes ! > Ngoc-Tien > > ----- Mail original ----- > De: "thibautjombart" > ?: "Ngoc-Tien Tran" > Cc: "adegenet-forum" > Envoy?: Lundi 11 Septembre 2017 14:05:46 > Objet: Re: [adegenet-forum] Loci selection for SNP data > > Hello, > > there is no tool specifically dedicated to this, but you could run a PCA, > retain axes corresponding to the xxx% of variance, and then keep only the > alleles contributing most to these axes. Here's a quick example: > > > data(sim2pop) > > pca1 <- dudi.pca(tab(sim2pop), scannf = FALSE, nf = 3) > > contrib <- pca1$c1^2 > > which(contrib > 0.01, TRUE) # contrib > 1% > row col > L02.06 21 1 > L04.02 47 1 > L05.02 60 1 > L05.09 67 1 > L06.6 74 1 > L07.7 84 1 > L08.07 91 1 > L09.8 102 1 > L09.9 103 1 > L10.01 104 1 > L10.12 115 1 > > [...] > > Best > Thibaut > > > -- > Dr Thibaut Jombart > Lecturer, Department of Infectious Disease Epidemiology, Imperial College > London > Head of RECON: repidemicsconsortium.org > WHO Consultant - outbreak analysis > sites.google.com/site/thibautjombart/ > Twitter: @TeebzR > +44(0)20 7594 3658 > > On 7 August 2017 at 13:45, Ngoc-Tien Tran > wrote: > > > Hello, > > I'm working with SNP data with a large number of loci. > > I would like to know if there is a tool in the package "adegenet" > allowing > > me to select just one smaller set of loci without loosing information. > > > > Thanks for your help ! > > Best wishes, > > Ngoc-Tien > > > > > > -- > > -------------------------------------------------------------- > > TRAN Ngoc-Tien > > D?partement G?nie Math?matiques/Analyse de Risque et Optimisation et > > Mod?lisation. > > INSA-Rennes > > Tel : +33.(0)7.51.57.96.53 > > @mail : ngoc-tien.tran at insa-rennes.fr > > _______________________________________________ > > 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 > -- > -------------------------------------------------------------- > TRAN Ngoc-Tien > D?partement G?nie Math?matiques/Analyse de Risque et Optimisation et > Mod?lisation. > INSA-Rennes > Tel : +33.(0)7.51.57.96.53 > @mail : ngoc-tien.tran at insa-rennes.fr > -------------- next part -------------- An HTML attachment was scrubbed... URL: From Mark.Coulson.ic at uhi.ac.uk Mon Sep 25 15:58:28 2017 From: Mark.Coulson.ic at uhi.ac.uk (Mark Coulson) Date: Mon, 25 Sep 2017 13:58:28 +0000 Subject: [adegenet-forum] xvalDapc and group prediction accuracy Message-ID: Hello, I am using xvalDapc on a dataset and it produces a nice arc-shape with the optimal number of PCs retained at 20. However, if I run the dapc function keeping 20 PCs, I get the warning that '#PCs may be too large (>N/3) and that results may be unstable' How should I interpret this given the choice to use 20 PCs was through the double-cross validation procedure? Thanks, Mark Inverness College UHI, a partner in the University of the Highlands and Islands www.inverness.uhi.ac.uk Board of Management of Inverness College (known as Inverness College UHI), Scottish Charity No SC021197. -------------- next part -------------- An HTML attachment was scrubbed... URL: From thibautjombart at gmail.com Wed Sep 27 12:18:40 2017 From: thibautjombart at gmail.com (Thibaut Jombart) Date: Wed, 27 Sep 2017 12:18:40 +0200 Subject: [adegenet-forum] xvalDapc and group prediction accuracy In-Reply-To: References: Message-ID: Hi there ignore the warning - it is ad-hoc, and should be removed. Best Thibaut -- Dr Thibaut Jombart Lecturer, Department of Infectious Disease Epidemiology, Imperial College London Head of RECON: repidemicsconsortium.org WHO Consultant - outbreak analysis sites.google.com/site/thibautjombart/ Twitter: @TeebzR +44(0)20 7594 3658 On 25 September 2017 at 15:58, Mark Coulson wrote: > Hello, > > > > I am using xvalDapc on a dataset and it produces a nice arc-shape with the > optimal number of PCs retained at 20. However, if I run the dapc function > keeping 20 PCs, I get the warning that > > > > ?#PCs may be too large (>N/3) and that results may be unstable? > > > > How should I interpret this given the choice to use 20 PCs was through the > double-cross validation procedure? > > > > Thanks, > > Mark > > Inverness College UHI, a partner in the University of the Highlands and > Islands www.inverness.uhi.ac.uk Board of Management of Inverness College > (known as Inverness College UHI), Scottish Charity No SC021197. > > _______________________________________________ > 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 From sand.ludwig at gmail.com Tue Sep 12 15:15:55 2017 From: sand.ludwig at gmail.com (Sandra Ludwig) Date: Tue, 12 Sep 2017 13:15:55 -0000 Subject: [adegenet-forum] Identification of mixture individuals Message-ID: Hello, I`m working with 8 loci of msats in several fish populations from Brazil, and during the analysis I'm using adegenet package to identify clusters (k) and also to estimate migration pattern between the populations. So...I'm wondering if I can assume that the individuals with mixed genotypes (see the output below) could be migrants individuals that were sampled in distinct populations? [image: Imagem inline 1] Thanks for your help!! Sandra ------ Sandra Ludwig Substitute Profesor of Zoology Laboratory of Biotecnology and Molecular Markers - E3 179 Universidade Federal de Minas Gerais, Belo Horizonte, Brazil Phone: (+55031) 97101-3052 s andraludwig at ufmg.br http://lattes.cnpq.br/6627603237397917 -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: image.png Type: image/png Size: 188068 bytes Desc: not available URL: