I'm working with capscale and permutest for the first time, and having
trouble getting statistical analyses for more than one constraining
variable. I've read the documentation, but setting first=FALSE or using
by="axis" doesn't seem to be helping. capscale seems to be fine, I
receive
output for more than one constrained axis. What am I doing wrong?
capscale.Nrem.results<-capscale(as.dist(qiime.data$distmat)~
N+rem+N*rem+Condition(dateFac), factor.frame)
capscale.Nrem.results
Inertia Proportion Rank
Total 1.454538
Real Total 1.459802 1.000000
Conditional 0.117117 0.080228 1
Constrained 0.386228 0.264576 3
Unconstrained 0.956457 0.655197 22
Imaginary -0.005264 2
Inertia is squared Unknown distance
Eigenvalues for constrained axes:
CAP1 CAP2 CAP3
0.29869 0.05395 0.03359
Eigenvalues for unconstrained axes:
MDS1 MDS2 MDS3 MDS4 MDS5 MDS6 MDS7 MDS8
0.27719 0.13725 0.11048 0.06691 0.05551 0.04940 0.03892 0.03468
(Showed only 8 of all 22 unconstrained eigenvalues)
sig.Nrem <- permutest(capscale.Nrem.results,permutations=999,
by="margin",
model="direct",first=FALSE)
sig.Nrem
Permutation test for capscale
Call: capscale(formula = as.dist(qiime.data$distmat) ~ N + rem +
Condition(dateFac) + N:rem, data factor.frame)
Permutation test for all constrained eigenvalues
Pseudo-F: 2.961281 (with 3, 22 Degrees of Freedom)
Significance: 0.001
Based on 999 permutations under direct model.
[[alternative HTML version deleted]]
