I am relatively new to setting up statistical analyses on R.
I am running a phylogenetic least squares regression for multiple observations per species per seasonality. (Example dataframe below). In all, there are 6 species measured (only one shown). Five individuals were measured for a trait values (CN shown) in three periods (Spring, Summer, Autumn).
| Tree_Species | Period | CN |
|---|---|---|
| TRE1 | Spring | 11.847 |
| TRE1 | Spring | 11.508 |
| TRE1 | Spring | 7.378 |
| TRE1 | Spring | 9.768 |
| TRE1 | Spring | 9.205 |
| TRE1 | Summer | 0.953 |
| TRE1 | Summer | 1.053 |
| TRE1 | Summer | 1.340 |
| TRE1 | Summer | 1.128 |
| TRE1 | Summer | 1.273 |
| TRE1 | Autumn | 1.153 |
| TRE1 | Autumn | 1.173 |
| TRE1 | Autumn | 1.366 |
| TRE1 | Autumn | 1.015 |
| TRE1 | Autumn | 1.108 |
In this example, I am trying to find whether seasonality (Period) has an affect on the measured trait (CN) given the species (Tree_Species).However, when I try to create a model while grouping the observations by species with the phylogeny (under brownian motion), I am recovering AIC, BIC, and log-likelihood values of infinity. The output (given by summary function) most likely means there is an issue in the way I am inputting the data, given the phylogeny.
The code that I performed on R is shown below with the libraries used.
library(phytools)
library(phylotools)
library(nlme)
library(piecewiseSEM)
library(picante)
library(geiger)
library(MuMIn)
library(bbmle)
traits <- read.delim("clipboard", as.is=FALSE) #import data
phy <-read.tree("mytimetree_strict_szn.tre") #import phylogeny, 6 tips. Ultrametric tree.
traits_grouped <- groupedData(CN ~ Period | Tree_species, data=traits, FUN=mean) #group the observations by species
bm.CN.season.gls <- gls(CN ~ Period, data=traits_grouped, correlation=corBrownian(1,phy, form= ~Tree_species))
summary(bm.CN.season.gls)
Generalized least squares fit by REML
Model: CN ~ Period
Data: traits_grouped
AIC BIC logLik
-Inf -Inf Inf
Correlation Structure: corBrownian
Formula: ~Tree_species
Parameter estimate(s):
numeric(0)
Coefficients:
Value Std.Error t-value p-value
(Intercept) 21.715098 3.123492 6.952186 0.0000
PeriodSpring -10.566643 2.485146 -4.251920 0.0001
PeriodSummer -1.847364 1.667929 -1.107580 0.2711
Correlation:
(Intr) PrdSpr
PeriodSpring -0.866
PeriodSummer -0.661 0.769
Standardized residuals:
Min Q1 Med Q3 Max
-3.36080010 -0.92884866 -0.08434883 0.46132229 3.30057650
Residual standard error: 2.472081
Degrees of freedom: 90 total; 87 residual
Would anyone be able to help me resolve this issue? I greatly appreciate the help and feedback!
