Re: [R-sig-phylo] Phylogenetic ANOVA
Correct! Thanks, Liam and Luke. Cheers, Ted Theodore Garland, Jr., Ph.D. Professor Department of Biology University of California Riverside, CA 92521 Phone: (951) 827-3524 = Ted's office (with answering machine) Phone: (951) 827-5724 = Ted's lab Phone: (951) 827-5903 = Dept. office Home Phone: (951) 328-0820 FAX: (951) 827-4286 = Dept. office Email: tgarl...@ucr.edu http://biology.ucr.edu/people/faculty/Garland.html List of all publications with PDF files: http://biology.ucr.edu/people/faculty/Garland/GarlandPublications.html Garland, T., Jr., and M. R. Rose, eds. 2009. Experimental evolution: concepts, methods, and applications of selection experiments. University of California Press, Berkeley, California. http://www.ucpress.edu/books/pages/10604.php Associate Director Network for Experimental Research on Evolution http://nere.bio.uci.edu/ (A University of California Multicampus Research Project) Original message Date: Mon, 26 Jul 2010 11:58:31 -0400 From: "Liam J. Revell" Subject: Re: [R-sig-phylo] Phylogenetic ANOVA To: Luke Harmon Cc: R phylo mailing list mailing list >Just to clarify the point - the null distribution for the test-statistic >(F) in this method is generated by Brownian motion simulation on the >phylogeny. The P-value of the ANOVA is thus obtained by comparing the >observed test-statistic to simulated test-statistics (for an arbitrarily >large number of simulations). The reference to this is: > >Garland, T. Jr., A. W. Dickerman, C. M. Janis, & J. A. Jones. 1993. >Phylogenetic analysis of covariance by computer simulation. Syst. Biol. >42: 265-292. (http://www.jstor.org/stable/2992464) > >Thus, if you have a test-statistic (F) more extreme then that obtained >for every last one of your simulated datasets, then the P-value will be >entirely determined by the number of simulations that are used (as Luke >says). This seems to be case for your data (not surprising given the >very large values for F that were obtained). > >- Liam > >Liam J. Revell >NESCent, Duke University >web: http://anolis.oeb.harvard.edu/~liam/ >NEW email: lrev...@nescent.org > > > >Luke Harmon wrote: >> Yes that's a direct result of the number of simulations - if all of the simulated F statistics are smaller than the test statistics, then you will get: >> >> p = 1/(n+1) where n is the number of simulated data sets. >> >> lh >> On Jul 26, 2010, at 8:44 AM, Alejandro Gonzalez V wrote: >> >> >>> Hello, >>> >>> Some colleagues and I are running some phylogenetic ANOVAS using the geiger package. In some of the analyses we get the same phylogentic p-value (very small p-value) even though the F-statistic differs between the two analyses, albeit it being relatively high in both instances. We were wondering why this arises, to get better grip on how the analysis works. We thought it may have to do with the randomizations to calculate the phylogenetic p-value. Or that the F-statistics are quite high... >>> Below are two examples : >>> >>> m11<-phy.anova(tree1,tmax,biozone,data.names=X,nsim=1000) >>> Standard ANOVA: >>> Analysis of Variance Table >>> >>> Response: td$data >>> Df Sum Sq Mean Sq F value Pr(>F) >>> group 1 967.96 967.96 155.88 3.057e-12 *** >>> Residuals 25 155.24 6.21 >>> --- >>> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 >>> >>> >>> Phylogenetic p-value: 0.000999001 >>> >>> >>> m12<-phy.anova(tree1,wt,biozone,data.names=X,nsim=1000) >>> Standard ANOVA: >>> Analysis of Variance Table >>> >>> Response: td$data >>> Df Sum Sq Mean Sq F value Pr(>F) >>> group 1 602.88 602.88 109.01 1.333e-10 *** >>> Residuals 25 138.26 5.53 >>> --- >>> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 >>> >>> >>> Phylogenetic p-value: 0.000999001 >>> >>> >>> Cheers, >>> >>> Alejandro >>> __ >>> >>> Alejandro Gonzalez Voyer >>> Post-doc >>> >>> NEW ADDRESS & NEW E-MAIL >>> >>> Estación Biológica de Doñana (CSIC) >>> Avenida Américo Vespucio s/n >>> 41092 Sevilla >>> Spain >>> >>> E-mail: alejan
Re: [R-sig-phylo] Phylogenetic ANOVA
Just to clarify the point - the null distribution for the test-statistic (F) in this method is generated by Brownian motion simulation on the phylogeny. The P-value of the ANOVA is thus obtained by comparing the observed test-statistic to simulated test-statistics (for an arbitrarily large number of simulations). The reference to this is: Garland, T. Jr., A. W. Dickerman, C. M. Janis, & J. A. Jones. 1993. Phylogenetic analysis of covariance by computer simulation. Syst. Biol. 42: 265-292. (http://www.jstor.org/stable/2992464) Thus, if you have a test-statistic (F) more extreme then that obtained for every last one of your simulated datasets, then the P-value will be entirely determined by the number of simulations that are used (as Luke says). This seems to be case for your data (not surprising given the very large values for F that were obtained). - Liam Liam J. Revell NESCent, Duke University web: http://anolis.oeb.harvard.edu/~liam/ NEW email: lrev...@nescent.org Luke Harmon wrote: Yes that's a direct result of the number of simulations - if all of the simulated F statistics are smaller than the test statistics, then you will get: p = 1/(n+1) where n is the number of simulated data sets. lh On Jul 26, 2010, at 8:44 AM, Alejandro Gonzalez V wrote: Hello, Some colleagues and I are running some phylogenetic ANOVAS using the geiger package. In some of the analyses we get the same phylogentic p-value (very small p-value) even though the F-statistic differs between the two analyses, albeit it being relatively high in both instances. We were wondering why this arises, to get better grip on how the analysis works. We thought it may have to do with the randomizations to calculate the phylogenetic p-value. Or that the F-statistics are quite high... Below are two examples : m11<-phy.anova(tree1,tmax,biozone,data.names=X,nsim=1000) Standard ANOVA: Analysis of Variance Table Response: td$data Df Sum Sq Mean Sq F valuePr(>F) group 1 967.96 967.96 155.88 3.057e-12 *** Residuals 25 155.246.21 --- Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 Phylogenetic p-value: 0.000999001 m12<-phy.anova(tree1,wt,biozone,data.names=X,nsim=1000) Standard ANOVA: Analysis of Variance Table Response: td$data Df Sum Sq Mean Sq F valuePr(>F) group 1 602.88 602.88 109.01 1.333e-10 *** Residuals 25 138.265.53 --- Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 Phylogenetic p-value: 0.000999001 Cheers, Alejandro __ Alejandro Gonzalez Voyer Post-doc NEW ADDRESS & NEW E-MAIL Estación Biológica de Doñana (CSIC) Avenida Américo Vespucio s/n 41092 Sevilla Spain E-mail: alejandro.gonza...@ebd.csic.es Tel: +34- 954 466700, ext 1749 Website (From my previous position): http://www.iee.uu.se/zooekol/default.php?type=personalpage&lang=en&id=146 [[alternative HTML version deleted]] ___ R-sig-phylo mailing list R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Luke Harmon Assistant Professor Biological Sciences University of Idaho 208-885-0346 lu...@uidaho.edu ___ R-sig-phylo mailing list R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo ___ R-sig-phylo mailing list R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Re: [R-sig-phylo] Phylogenetic ANOVA
Yes that's a direct result of the number of simulations - if all of the simulated F statistics are smaller than the test statistics, then you will get: p = 1/(n+1) where n is the number of simulated data sets. lh On Jul 26, 2010, at 8:44 AM, Alejandro Gonzalez V wrote: > Hello, > > Some colleagues and I are running some phylogenetic ANOVAS using the geiger > package. In some of the analyses we get the same phylogentic p-value (very > small p-value) even though the F-statistic differs between the two analyses, > albeit it being relatively high in both instances. We were wondering why this > arises, to get better grip on how the analysis works. We thought it may have > to do with the randomizations to calculate the phylogenetic p-value. Or that > the F-statistics are quite high... > Below are two examples : > > m11<-phy.anova(tree1,tmax,biozone,data.names=X,nsim=1000) > Standard ANOVA: > Analysis of Variance Table > > Response: td$data > Df Sum Sq Mean Sq F valuePr(>F) > group 1 967.96 967.96 155.88 3.057e-12 *** > Residuals 25 155.246.21 > --- > Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 > > > Phylogenetic p-value: 0.000999001 > > > m12<-phy.anova(tree1,wt,biozone,data.names=X,nsim=1000) > Standard ANOVA: > Analysis of Variance Table > > Response: td$data > Df Sum Sq Mean Sq F valuePr(>F) > group 1 602.88 602.88 109.01 1.333e-10 *** > Residuals 25 138.265.53 > --- > Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 > > > Phylogenetic p-value: 0.000999001 > > > Cheers, > > Alejandro > __ > > Alejandro Gonzalez Voyer > Post-doc > > NEW ADDRESS & NEW E-MAIL > > Estación Biológica de Doñana (CSIC) > Avenida Américo Vespucio s/n > 41092 Sevilla > Spain > > E-mail: alejandro.gonza...@ebd.csic.es > > Tel: +34- 954 466700, ext 1749 > > Website (From my previous position): > http://www.iee.uu.se/zooekol/default.php?type=personalpage&lang=en&id=146 > > > > > > > > > > [[alternative HTML version deleted]] > > ___ > R-sig-phylo mailing list > R-sig-phylo@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Luke Harmon Assistant Professor Biological Sciences University of Idaho 208-885-0346 lu...@uidaho.edu ___ R-sig-phylo mailing list R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo