Off-topic, but I wanted to comment on Anthony's wariness about commenting on R-sig lists...
Yes, it can certainly be very difficult to give advice that is tailored specifically to the problem of a particular worker on R-sig lists, particularly as one can't just tell the other person to open their data files and show them to you. However, answering a question on R-sig lists (regardless of whether it is the right answer, but what answer remains right forever anyway?) records that answer virtually forever, for future posterity. This means the answer can be found by any future individuals who encounter this problem, via a simple google search. Since I first found R-sig-phylo in early 2010 (five years ago!), I have made almost constant use of the knowledge base contained with R-sig-phylo's archive. It doesn't mean the information is always right, or always the best answers for the actual person who asked originally, but it certainly helps point out the right line of thinking or the right literature to investigate. StackExchange discussion archives have become just as valuable (but there appears to be very little R phylo discussion over there). It feels like the amount of discussion on the list has dropped off slowly over the last year, and as phylogenetics in R seems as widely used as ever, I have often wondered if this reflects that most issues people may run into are now more easily solved with google searches leading to the R-sig-phylo archives. Anyway, I just hope that this element is not forgotten about why I think replying to question on R-sig-phylo remains a valuable contribution to our community! Cheers, -Dave Bapst On Mon, Mar 2, 2015 at 7:58 AM, Anthony Ives <[email protected]> wrote: > Simon and Ben, > > Of course, sample size of 8 is going to be an issue in almost any analysis. > But sometimes that is all the data there are. > > Incidentally, this exchange reminded me that I’m still wary of making > comments on r-sig. If somebody comes into my office, I have the time to > discuss with them their data, so I can learn more about it. Then I feel I can > at least make informed recommendations for analyses — they might still be > badly wrong recommendations, but at least they are informed. I’m still > uncomfortable about making suggestions on r-sig, when I don’t really have > full information, or the time to think. Therefore, the few comments I’ve made > have been very general about methods, rather than specific about data sets. > > I think this is just a matter of me waking up to the 21st century. I do like > the idea of crowdsourcing; I just need to get comfortable with it. > > Cheers, Tony > > > Anthony Ives > Department of Zoology > 459 Birge Hall (4th floor, E end of bldg) > UW-Madison > Madison, WI 53706 > 608-262-1519 > >> On Mar 1, 2015, at 10:53 PM, Simon Blomberg <[email protected]> wrote: >> >> Hi Ben, >> >> Yes, you would have to assume constant variance across species to use N=24. >> I think that is the only option. But given that biological data often has a >> positive mean-variance relationship, again I'm dubious about the exercise. >> YMMV, however! >> >> Cheers, >> >> Simon. >> >> Simon Blomberg, BSc (Hons), PhD, MAppStat, AStat. >> Senior Lecturer and Consultant Statistician >> School of Biological Sciences >> The University of Queensland >> St. Lucia Queensland 4072 >> Australia >> >> T: +61 7 3365 2506 >> email: S.Blomberg1_at_uq.edu.au >> http://www.evolutionarystatistics.org >> >> Policies: >> >> 1. I will NOT analyse your data for you. >> 2. Your deadline is your problem >> >> Basically, I'm not interested in doing research and I never have been. I'm >> interested in understanding, which is quite a different thing. - David >> Blackwell >> >> ________________________________________ >> From: R-sig-phylo [[email protected]] on behalf of Ben >> Bolker [[email protected]] >> Sent: Monday, March 02, 2015 2:49 PM >> To: [email protected] >> Subject: Re: [R-sig-phylo] phytools - evaluating significance of pgls.Ives >> >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> On 15-03-01 11:40 PM, Simon Blomberg wrote: >>> Am I missing something? The OP only has 8 species in the data set. >>> I wouldn't put much store in fancy PCM modelling based on such a >>> small data set. And 3 individuals per species is not enough for a >>> good estimate of the within-species variance. >>> >>> Simon. >> >> Agree wholeheartedly with the first point -- but for the second, >> isn't 24 rather than 8 the relevant number for estimating >> within-species variance (since presumably we are assuming the same >> variance within every species, thus we can effectively pool >> within-species variation \across species for this purpose) ? >> >>> >>> Simon Blomberg, BSc (Hons), PhD, MAppStat, AStat. Senior Lecturer >>> and Consultant Statistician School of Biological Sciences The >>> University of Queensland St. Lucia Queensland 4072 Australia >>> >>> T: +61 7 3365 2506 email: S.Blomberg1_at_uq.edu.au >>> http://www.evolutionarystatistics.org >>> >>> Policies: >>> >>> 1. I will NOT analyse your data for you. 2. Your deadline is >>> your problem >>> >>> Basically, I'm not interested in doing research and I never have >>> been. I'm interested in understanding, which is quite a different >>> thing. - David Blackwell >>> >>> ________________________________________ From: R-sig-phylo >>> [[email protected]] on behalf of Anthony R Ives >>> [[email protected]] Sent: Monday, March 02, 2015 2:14 PM To: Andrea >>> Berardi Cc: [email protected] Subject: Re: [R-sig-phylo] >>> phytools - evaluating significance of pgls.Ives >>> >>> Andrea, >>> >>> I second Liam’s recommendation to use a LRT. >>> >>> For measurement error, the latest code I have in matlab is >>> MERegPHYSIGv2.m, which does both measurement error and an OU or >>> Pagel-lambda transform (see Johnson, M. T. J., A. R. Ives, J. >>> Ahern, and J. P. Salminen. 2014. Macroevolution of plant defenses >>> against herbivores in the evening primroses. New Phytologist >>> 203:267-279). Measurement-error models are always going to have >>> difficulties at parameter boundaries; for example, if the assumed >>> measurement error is large, it can exceed the observed variation in >>> the data, which of course causes problems (statistical and >>> logical). >>> >>> In MERegPHYSIGv2.m, I did a round or two of simulated annealing >>> first, before polishing the results with a Nelder-Mead optimizer. >>> It seems like you could do the same with Liam’s code pretty easily >>> by changing the method of optimization (using edit()). Before >>> doing this, thought, I would take a careful look at your data and >>> your estimates of measurement error. An easy diagnostic is to start >>> with 10% of your estimated measurement standard errors and then >>> increase slowly to 100%. When I have done this, I’ve been able to >>> see problems when parameter values go awry. It is not a fail-safe >>> diagnostic in any way, but it can help. >>> >>> Cheers, Tony >>> >>> >>> >>> Anthony Ragnar Ives Department of Zoology UW-Madison Madison, WI >>> 53706 608-262-1519 >>> >>>> On Mar 1, 2015, at 9:42 PM, Liam J. Revell <[email protected]> >>>> wrote: >>>> >>>> Hi Andrea. >>>> >>>> This is not presently implemented, but since this is a >>>> likelihood method it would be straightforward to constrain to a >>>> slope of zero and then do a LR test. This would be probably be >>>> the easiest way to test a hypothesis about the regression. >>>> >>>> That being said, as noted in the function documentation, some >>>> problems have been reported with the optimization algorithm for >>>> this model, which is simple and thus may fail to find the ML >>>> solution. Consequently, I would encourage you to look for other >>>> implementations of the method so that you can be confident in >>>> your result. I'm not aware of one in R at this time. >>>> >>>> All the best, Liam >>>> >>>> Liam J. Revell, Assistant Professor of Biology University of >>>> Massachusetts Boston web: http://faculty.umb.edu/liam.revell/ >>>> email: [email protected] blog: http://blog.phytools.org >>>> >>>> On 3/1/2015 10:31 PM, Andrea Berardi wrote: >>>>> Hi all, >>>>> >>>>> I'm just learning how to do PGLS analyses, and I'm looking for >>>>> advice on how to evaluate the significance of the regression >>>>> fit using pgls.Ives in the phytools package. I'm using this >>>>> function because it incorporates sampling error of species >>>>> means, and my data has about 3 individuals per species, with 8 >>>>> species. My goal is to test whether a flower trait predicts the >>>>> leaf trait, while controlling for shared ancestry. Here is the >>>>> output from pgls.Ives: >>>>> >>>>>> fit <- pgls.Ives(Tree, Flower_trait, Leaf_trait) fit >>>>> $beta [1] 96.3963098 0.1292656 >>>>> >>>>> $sig2x [1] 22218901073 >>>>> >>>>> $sig2y [1] 23027587 >>>>> >>>>> $a [1] -10063.150 -1204.422 >>>>> >>>>> $logL [1] -158.2337 >>>>> >>>>> $convergence [1] 0 >>>>> >>>>> $message [1] "CONVERGENCE: REL_REDUCTION_OF_F <= FACTR*EPSMCH" >>>>> >>>>> I am also running pgls on species averages for the traits >>>>> using the gls function in nlme and the corBrownian and >>>>> corMartins functions in ape. But, we are interested in >>>>> incorporating the within-species variation in our small >>>>> dataset. >>>>> >>>>> Any suggestions would be welcome! >>>>> >>>>> Thanks for your help, Andrea >>>>> >>>>> ~~ Andrea Berardi, PhD Postdoctoral Researcher, Smith Lab >>>>> EBIO, University of Colorado-Boulder >>>>> >>>>> _______________________________________________ R-sig-phylo >>>>> mailing list - [email protected] >>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable >>>>> archive at >>>>> http://www.mail-archive.com/[email protected]/ >>>>> >>>> >>>> _______________________________________________ R-sig-phylo >>>> mailing list - [email protected] >>>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable >>>> archive at >>>> http://www.mail-archive.com/[email protected]/ >>> >>> _______________________________________________ R-sig-phylo >>> mailing list - [email protected] >>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable >>> archive at http://www.mail-archive.com/[email protected]/ >>> >>> _______________________________________________ R-sig-phylo >>> mailing list - [email protected] >>> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable >>> archive at http://www.mail-archive.com/[email protected]/ >>> >> >> -----BEGIN PGP SIGNATURE----- >> Version: GnuPG v1.4.11 (GNU/Linux) >> >> iQEcBAEBAgAGBQJU8+vwAAoJEOCV5YRblxUHCrIIALM+JBir5Bieh4FpqwU7XiVh >> kekQkoMxXmAmYigqyXxFAuUqrvkYk+kNZi+Z4EzqU3LOASRY32oFO1ZReMTEvQNf >> 8R2JtwaD8Sfr/+fb8juJalYoKnFfNIqxZmuKpRsobJmaiKxd9bD22zKLpGMZDVDO >> a8nssrBkoZZvvtegq7EaMGiGMEDs8TfynzXU2l1TDBJ45IyLTvwAP5SXrfstPzbq >> HyuKpTFXVs1+9GmmiOB3NuPzKtXwIg2dG8ahN6FhCPuoCkSI4XFx2k581TnsAh3o >> wUqyTGhKYuRhJVyjHtOYlSDRHoiE+5yN9fwlNNUnfTQnVdP7TTwyB+Xg+jtiQAQ= >> =tQqZ >> -----END PGP SIGNATURE----- >> >> _______________________________________________ >> R-sig-phylo mailing list - [email protected] >> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo >> Searchable archive at http://www.mail-archive.com/[email protected]/ >> >> _______________________________________________ >> R-sig-phylo mailing list - [email protected] >> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo >> Searchable archive at http://www.mail-archive.com/[email protected]/ > > _______________________________________________ > R-sig-phylo mailing list - [email protected] > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo > Searchable archive at http://www.mail-archive.com/[email protected]/ -- David W. Bapst, PhD Adjunct Asst. Professor, Geology and Geol. Eng. South Dakota School of Mines and Technology 501 E. St. Joseph Rapid City, SD 57701 http://webpages.sdsmt.edu/~dbapst/ http://cran.r-project.org/web/packages/paleotree/index.html _______________________________________________ R-sig-phylo mailing list - [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/[email protected]/
