Re: [R-sig-phylo] ape chronos function returning pLL of -1E100
Hi John, It looks like that the constraints on the dates with the 4 calibrations are such that the penalized likelihood cannot be calculated. There are several occasions when a value of -1e100 is returned (non finite value, negative branch lenghts, ...) so it's hard to see what is the problem here. Maybe it's better we discuss about this off-list. Best, Emmanuel Le 13/12/2018 à 04:05, John S Denton a écrit : As an update to my question--when I try to calibrate the tree itself using the default single root calibration and default 1/1 min/max ages, the function appears to run fine. When I add in my calibration table, the function exits with the -Inf LL. The tree is rooted. There are no duplicated nodes with conflicting ages, and R does not give errors about the formatting of the calibration table. ~John From: R-sig-phylo on behalf of John S Denton Sent: Wednesday, December 12, 2018 1:20:20 PM To: r-sig-phylo@r-project.org Subject: [R-sig-phylo] ape chronos function returning pLL of -1E100 I'm trying to use chronos (ape version 5.2) to scale a tree of ~800 tips, with branch lengths derived from a RAxML analysis. I'm using 4 outgroup calibrations. When I run the analysis using t.og <- chronos(tree, lambda = 1, model = "relaxed", quiet = FALSE, calibration = dd.out) I get the following returned, in almost 5 seconds: Setting initial dates... Fitting in progress... get a first set of estimates Penalised log-lik = -1e+100 Optimising rates... dates... -1e+100 Done. a look at attributes (attr(t.og, "message")) gives a "relative convergence (4)." The pLL, however, is -Inf. I have tried setting chronos.control to different values, with no success. It looks like it may be an issue with the optimization function inside chronos. Are there additional constraints/code I can apply to make it optimize without (apparently) simply running to the boundary? Thanks, ~John ___ 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] ape chronos function returning pLL of -1E100
As an update to my question--when I try to calibrate the tree itself using the default single root calibration and default 1/1 min/max ages, the function appears to run fine. When I add in my calibration table, the function exits with the -Inf LL. The tree is rooted. There are no duplicated nodes with conflicting ages, and R does not give errors about the formatting of the calibration table. ~John From: R-sig-phylo on behalf of John S Denton Sent: Wednesday, December 12, 2018 1:20:20 PM To: r-sig-phylo@r-project.org Subject: [R-sig-phylo] ape chronos function returning pLL of -1E100 I'm trying to use chronos (ape version 5.2) to scale a tree of ~800 tips, with branch lengths derived from a RAxML analysis. I'm using 4 outgroup calibrations. When I run the analysis using t.og <- chronos(tree, lambda = 1, model = "relaxed", quiet = FALSE, calibration = dd.out) I get the following returned, in almost 5 seconds: Setting initial dates... Fitting in progress... get a first set of estimates Penalised log-lik = -1e+100 Optimising rates... dates... -1e+100 Done. a look at attributes (attr(t.og, "message")) gives a "relative convergence (4)." The pLL, however, is -Inf. I have tried setting chronos.control to different values, with no success. It looks like it may be an issue with the optimization function inside chronos. Are there additional constraints/code I can apply to make it optimize without (apparently) simply running to the boundary? Thanks, ~John [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylodata=01%7C01%7Cjdenton%40amnh.org%7C74f64ad9321e496dcbf208d6605e850e%7Cbe0003e8c6b9496883aeb34586974b76%7C0sdata=LEcC13%2BvcHJ3NY8su7oRd03U53SfnJGehwaTUYKblLw%3Dreserved=0 Searchable archive at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2Fdata=01%7C01%7Cjdenton%40amnh.org%7C74f64ad9321e496dcbf208d6605e850e%7Cbe0003e8c6b9496883aeb34586974b76%7C0sdata=lkx0ZcpMnQgumw2CvwMrFzKztku1bMgK%2Fi8IXmdwuWU%3Dreserved=0 [[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] ape chronos function returning pLL of -1E100
I'm trying to use chronos (ape version 5.2) to scale a tree of ~800 tips, with branch lengths derived from a RAxML analysis. I'm using 4 outgroup calibrations. When I run the analysis using t.og <- chronos(tree, lambda = 1, model = "relaxed", quiet = FALSE, calibration = dd.out) I get the following returned, in almost 5 seconds: Setting initial dates... Fitting in progress... get a first set of estimates Penalised log-lik = -1e+100 Optimising rates... dates... -1e+100 Done. a look at attributes (attr(t.og, "message")) gives a "relative convergence (4)." The pLL, however, is -Inf. I have tried setting chronos.control to different values, with no success. It looks like it may be an issue with the optimization function inside chronos. Are there additional constraints/code I can apply to make it optimize without (apparently) simply running to the boundary? Thanks, ~John [[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/