Re: [R-sig-phylo] Error in diversitree with MuSSE models
Hi Kevin, Guessing here at possible solutions: Have a close look at: start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)] You may have negative or NA values in there. You could try running find.mle with lower=0, but that should not be needed. I suspect that you got the error while the find.mle function was evaluating the starting point. If that's the case try: lik.allcon(start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) (or more readably: x.init - start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) lik.allcon(x.init) to check that you get the same error. If you do, then something is up with the constrained function. Check that after typing 'constrain' you see environment: namespace:diversitree not environment: namespace:geiger otherwise, load diversitree after geiger, or upgrade to the current geiger which solves a number of incompatibilities with diversitree. FWIW, I don't normally recommend constraining like that: I usually constrain all rates onto a single rate, e.g. ..., lambda2 ~ lambda1, lambda3 ~ lambda1, lambda4 ~ lambda1, ... but you should get an error from constrain() if it can't resolve your constraints. If all that fails, can you post a minimal working example (data and script that generates the error) perhaps as a gist and I can have a further poke about. Cheers, Rich On 3 April 2014 00:23, Arbuckle, Kevin k.arbuc...@liverpool.ac.uk wrote: Hi Nick, it's running (or trying to) on diversitree v0.9-7 Kev From: Nicholas Crouch [ncro...@uic.edu] Sent: 01 April 2014 17:01 To: Arbuckle, Kevin Subject: Re: [R-sig-phylo] Error in diversitree with MuSSE models Kev, What version of diversitree are you running? Nick On Apr 1, 2014 10:50 AM, Arbuckle, Kevin k.arbuc...@liverpool.ac.ukmailto:k.arbuc...@liverpool.ac.uk wrote: Hi, I am trying to run a set of MuSSE models in diversitree but getting error messages that I can't seem to figure out (fortunately the models are only an adjunct to my main premise, not particularly vital). The trait has 4 levels and the full model works fine, with the following code:- lik-make.musse(contree,col,k=4,sampling.f=0.33) start-starting.point.musse(contree,k=4) fit-find.mle(lik,start) However, when I then try and run a fully constrained model (with only a single speciation rate, extinction rate, and transition rate) with the following code:- lik.allcon-constrain(lik,lambda1~lambda2,lambda2~lambda3,lambda3~lambda4,mu1~mu2,mu2~mu3,mu3~mu4,q12~q13,q13~q14,q14~q21,q21~q23,q23~q24,q24~q31,q31~q32,q32~q34,q34~q41,q41~q42,q42~q43) fit.allcon-find.mle(lik.allcon,start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) I get the following error message:- Error in check.pars.nonnegative(pars, k * (k + 1)) : Parameters must be non-negative and finite I don't quite understand why or how to fix this. I have previously run BiSSE models (including with constraints) on the same dataset (just with different coding of some levels which made the trait binary) and had no problems. After looking through the diversitree manual I have tried specifying/changing the fail.value and method arguments in find.mle, but I get exactly the same message. Any help or ideas would be greatly appreciated. I presume it is something introduced via the constraints since the full model runs fine, but other than that I am none the wiser. Thanks, Kev [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.orgmailto:R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
Re: [R-sig-phylo] Error in diversitree with MuSSE models (Arbuckle, Kevin)
Hi Kev, Your constrain looks fine but I think you should double check your index for the 'find.mle' function in this line: fit.allcon - find.mle(lik.allcon,start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) 'start' is a vector with a number of elements equal to the number of parameters in your full model. To find the MLE of your constrained likelihood function you need to point to the free parameters in 'start'. Try this, check if start has only lambda1,mu1,q12: fit.allcon - find.mle(lik.allcon,start[c(1,6,12)]) Good luck *Daniel Caetano Silva* Ph.D. Student Harmon Lab http://caetanods.weebly.com/ http://www.webpages.uidaho.edu/~lukeh/ University of Idaho - Department of Biological Sciences Moscow, Idaho 83844-3051 USA On 04/03/2014 03:00 AM, r-sig-phylo-requ...@r-project.org wrote: Send R-sig-phylo mailing list submissions to r-sig-phylo@r-project.org To subscribe or unsubscribe via the World Wide Web, visit https://stat.ethz.ch/mailman/listinfo/r-sig-phylo or, via email, send a message with subject or body 'help' to r-sig-phylo-requ...@r-project.org You can reach the person managing the list at r-sig-phylo-ow...@r-project.org When replying, please edit your Subject line so it is more specific than Re: Contents of R-sig-phylo digest... Today's Topics: 1. Re: Error in diversitree with MuSSE models (Arbuckle, Kevin) -- Message: 1 Date: Wed, 2 Apr 2014 13:23:44 + From: Arbuckle, Kevin k.arbuc...@liverpool.ac.uk To: Nicholas Crouch ncro...@uic.edu Cc: r-sig-phylo@r-project.org r-sig-phylo@r-project.org Subject: Re: [R-sig-phylo] Error in diversitree with MuSSE models Message-ID: eb15200a53ef104a85d6d8e071cc0edcb2131...@bhexmbx2.livad.liv.ac.uk Content-Type: text/plain Hi Nick, it's running (or trying to) on diversitree v0.9-7 Kev From: Nicholas Crouch [ncro...@uic.edu] Sent: 01 April 2014 17:01 To: Arbuckle, Kevin Subject: Re: [R-sig-phylo] Error in diversitree with MuSSE models Kev, What version of diversitree are you running? Nick On Apr 1, 2014 10:50 AM, Arbuckle, Kevin k.arbuc...@liverpool.ac.ukmailto:k.arbuc...@liverpool.ac.uk wrote: Hi, I am trying to run a set of MuSSE models in diversitree but getting error messages that I can't seem to figure out (fortunately the models are only an adjunct to my main premise, not particularly vital). The trait has 4 levels and the full model works fine, with the following code:- lik-make.musse(contree,col,k=4,sampling.f=0.33) start-starting.point.musse(contree,k=4) fit-find.mle(lik,start) However, when I then try and run a fully constrained model (with only a single speciation rate, extinction rate, and transition rate) with the following code:- lik.allcon-constrain(lik,lambda1~lambda2,lambda2~lambda3,lambda3~lambda4,mu1~mu2,mu2~mu3,mu3~mu4,q12~q13,q13~q14,q14~q21,q21~q23,q23~q24,q24~q31,q31~q32,q32~q34,q34~q41,q41~q42,q42~q43) fit.allcon-find.mle(lik.allcon,start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) I get the following error message:- Error in check.pars.nonnegative(pars, k * (k + 1)) : Parameters must be non-negative and finite I don't quite understand why or how to fix this. I have previously run BiSSE models (including with constraints) on the same dataset (just with different coding of some levels which made the trait binary) and had no problems. After looking through the diversitree manual I have tried specifying/changing the fail.value and method arguments in find.mle, but I get exactly the same message. Any help or ideas would be greatly appreciated. I presume it is something introduced via the constraints since the full model runs fine, but other than that I am none the wiser. Thanks, Kev [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.orgmailto:R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ [[alternative HTML version deleted]] -- ___ R-sig-phylo mailing list R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo End of R-sig-phylo Digest, Vol 75, Issue 3 ** [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
Re: [R-sig-phylo] Error in diversitree with MuSSE models
Hi Nick, it's running (or trying to) on diversitree v0.9-7 Kev From: Nicholas Crouch [ncro...@uic.edu] Sent: 01 April 2014 17:01 To: Arbuckle, Kevin Subject: Re: [R-sig-phylo] Error in diversitree with MuSSE models Kev, What version of diversitree are you running? Nick On Apr 1, 2014 10:50 AM, Arbuckle, Kevin k.arbuc...@liverpool.ac.ukmailto:k.arbuc...@liverpool.ac.uk wrote: Hi, I am trying to run a set of MuSSE models in diversitree but getting error messages that I can't seem to figure out (fortunately the models are only an adjunct to my main premise, not particularly vital). The trait has 4 levels and the full model works fine, with the following code:- lik-make.musse(contree,col,k=4,sampling.f=0.33) start-starting.point.musse(contree,k=4) fit-find.mle(lik,start) However, when I then try and run a fully constrained model (with only a single speciation rate, extinction rate, and transition rate) with the following code:- lik.allcon-constrain(lik,lambda1~lambda2,lambda2~lambda3,lambda3~lambda4,mu1~mu2,mu2~mu3,mu3~mu4,q12~q13,q13~q14,q14~q21,q21~q23,q23~q24,q24~q31,q31~q32,q32~q34,q34~q41,q41~q42,q42~q43) fit.allcon-find.mle(lik.allcon,start[-c(1,2,3,5,6,7,9,10,11,12,13,14,15,16,17,18,19)]) I get the following error message:- Error in check.pars.nonnegative(pars, k * (k + 1)) : Parameters must be non-negative and finite I don't quite understand why or how to fix this. I have previously run BiSSE models (including with constraints) on the same dataset (just with different coding of some levels which made the trait binary) and had no problems. After looking through the diversitree manual I have tried specifying/changing the fail.value and method arguments in find.mle, but I get exactly the same message. Any help or ideas would be greatly appreciated. I presume it is something introduced via the constraints since the full model runs fine, but other than that I am none the wiser. Thanks, Kev [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.orgmailto:R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
