tag:blogger.com,1999:blog-8499895524521663926.post3284004370323994368..comments2024-03-27T07:13:39.236-04:00Comments on Phylogenetic Tools for Comparative Biology: New phylogenetic ANOVA function with posthoc testsLiam Revellhttp://www.blogger.com/profile/04314686830842384151noreply@blogger.comBlogger16125tag:blogger.com,1999:blog-8499895524521663926.post-19455332567137501162022-03-03T10:20:58.949-05:002022-03-03T10:20:58.949-05:00Still having the same problem hereStill having the same problem hereAnonymoushttps://www.blogger.com/profile/00428694678473843238noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-69116465404211657792020-10-07T12:02:16.742-04:002020-10-07T12:02:16.742-04:00Hi I am actually having this same issue right now ...Hi I am actually having this same issue right now and would love to hear if there is a solution! Anonymoushttps://www.blogger.com/profile/03297347657704633815noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-35883719360145313122020-09-28T05:51:28.985-04:002020-09-28T05:51:28.985-04:00Hi Everyone,
Also still looking for a solution for...Hi Everyone,<br />Also still looking for a solution for this.<br />Thank you!Anonymoushttps://www.blogger.com/profile/10108686740555387423noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-65433275296103392982019-05-17T07:24:18.448-04:002019-05-17T07:24:18.448-04:00Hello, i've seen several people with this prob...Hello, i've seen several people with this problem but haven't seen a solution yet, is there one? Thank youAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-85977036274345678652017-03-20T01:00:41.116-04:002017-03-20T01:00:41.116-04:00Dear Liam,
First of all, thanks a lot for this bl...Dear Liam,<br /><br />First of all, thanks a lot for this blog. I am very new in R, and it has helped me a lot.<br /><br />I wonder if there is a way to include more than one individual per species/tip and run a two-way ANOVA (with post hoc tests) using phylANOVA. Specifically, I’m trying to test for interspecific differences in a trait with species and (e.g.) reproductive status as factors. <br /><br />Thanks,<br /><br />Henrique<br />Henrique Brazhttps://www.blogger.com/profile/02636482147864787595noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-18049027588239707482017-02-06T13:47:55.309-05:002017-02-06T13:47:55.309-05:00I would search for how to do a post-hoc test in nl...I would search for how to do a post-hoc test in nlme or with the function gls. This is not a phytools function, so I don't know - but I suspect that methods have been written for precisely this purpose. - LiamLiam Revellhttps://www.blogger.com/profile/04314686830842384151noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-73424567632894357842017-02-06T12:24:53.553-05:002017-02-06T12:24:53.553-05:00Liam, I searched your blog for notes on how to con...Liam, I searched your blog for notes on how to conduct a post hoc test for a 2-way ANOVA and couldn't find anything. I can get the gls code to run but still need a way to do contrasts. Thanks!Liz Schultheishttps://www.blogger.com/profile/14767660849264253169noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-40598745387213975232016-09-29T10:07:39.487-04:002016-09-29T10:07:39.487-04:00Is there a solution to this yet?Is there a solution to this yet?Anonymoushttps://www.blogger.com/profile/07578348663274168595noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-88494877773070251662015-09-24T14:58:38.471-04:002015-09-24T14:58:38.471-04:00Hi Liam,
Is there a way to perform a phylogenetic...Hi Liam,<br /><br />Is there a way to perform a phylogenetic ANCOVA?<br /><br />The following sample function doesn't work since the formula is required to be in the form data~group<br /><br />aov.phylo(data~groups + covariate, tree, nsim=50)Anonymoushttps://www.blogger.com/profile/06848445703593289579noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-81437078335110519612014-09-19T22:39:52.386-04:002014-09-19T22:39:52.386-04:00Hi Rick.
You might want to do this using gls in t...Hi Rick.<br /><br />You might want to do this using gls in the nlme package, i.e.:<br /><br />fit<-gls(y~x1+x2+x1*x2,correlation=corBrownian(1,tree))<br /><br />for instance. Search for gls on my blog or the web for more information.<br /><br />- LiamLiam Revellhttps://www.blogger.com/profile/04314686830842384151noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-21975278928761373662014-09-18T16:14:23.725-04:002014-09-18T16:14:23.725-04:00Hey Liam,
Is it possible to do a phylogentic 2-wa...Hey Liam,<br /><br />Is it possible to do a phylogentic 2-way ANOVA? I can't seem to get either aov.phylo or phylANOVA to work.<br /><br />Thanks,<br />Rick SimpsonAnonymoushttps://www.blogger.com/profile/01152860177603323342noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-60556016310099293732014-01-30T16:24:27.287-05:002014-01-30T16:24:27.287-05:00Hi Sanuel.
phylANOVA computes t-values for all pai...Hi Sanuel.<br />phylANOVA computes t-values for all pairwise comparisons (although it could also just use the raw difference in the mean - it wouldn't matter), then obtains p-values for each comparison via simulation. To correct for multiple tests it uses p.adjust and the default method ("holm") is the default, although this can be changed.<br />- LiamLiam Revellhttps://www.blogger.com/profile/04314686830842384151noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-1826773053827599612014-01-30T12:56:49.866-05:002014-01-30T12:56:49.866-05:00Hi Lian,
I have a simple question: what is the nam...Hi Lian,<br />I have a simple question: what is the name of your posthoc test? Is Tukey the default?<br />Thanks!<br />Samuelnoreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-23924049803555887372012-09-24T08:08:27.096-04:002012-09-24T08:08:27.096-04:00Hi Liam, thanks for the great code! I thought you ...Hi Liam, thanks for the great code! I thought you might be interested to know that the post-hoc tests fail when one of the groups is represented by a single case (it returns NA for all pairs, even ones with multiple cases per group). Not sure if there is an easy way around this. christoferhttps://www.blogger.com/profile/02615009080339789957noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-48581350425528382872011-08-28T12:06:58.740-04:002011-08-28T12:06:58.740-04:00Hi Roger. I think an implicit assumption is BM ev...Hi Roger. I think an implicit assumption is BM evolution of the residual error controlling for the group effect. (A strict BM assumption wouldn't make any sense because under BM there would be no tendency for species in the same group to have similar phenotypes.) I hope that is helpful, but I would encourage you to check out the primary literature on this method for more specifics. Best of luck, LiamLiam Revellhttps://www.blogger.com/profile/04314686830842384151noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-88362545271448002792011-08-25T00:22:19.036-04:002011-08-25T00:22:19.036-04:00Hi Liam, thanks for making this available; it look...Hi Liam, thanks for making this available; it looks very useful. Does it assume that the trait is evolving under BM evolution? I'd like to use it on some data that seem to best fit OU (having run fitContinuous on it; branch lengths of tree scaled using the method of Brusatte et al. 2008 using Graeme Lloyd's script). Is this a problem for phylANOVA?Roger Closehttps://www.blogger.com/profile/03485968561837748536noreply@blogger.com