[R-sig-eco] Problem in matrix contrasts using adonis function in vegan package

[ASANTOS]

Dear Members,

        I would like to comparing the default aov-table using adonis 
function in vegan package for the customized contrasts for can see what 
the residual and the total sum of squares for different models. But I 
maked any mistake when I create a design matrix of the contrasts for 
"avaliacao" object in my example, any Member can help me, Thanks!!

#Packages
require(vegan)

#Artificial data
colony<-as.factor(sort(rep(c("colony1", "colony2","colony3"), 100)))
y1 <- c(rnorm(100,1,0.1),rnorm(100,5,0.1),rnorm(100,3.5,0.1))
y2 <- c(rnorm(100,10,0.3),rnorm(100,11,0.6),rnorm(100,5,0.6))
y3 <- c(rnorm(100,10,2.3),rnorm(100,11,2.6),rnorm(100,11,2.5))
y4 <- c(rnorm(100,5,0.5),rnorm(100,7,0.5),rnorm(100,22,0.5))
y5 <- c(rnorm(100,11,0.1),rnorm(100,11,0.1),rnorm(100,11,0.1))


#Data frame
avaliacao <- as.factor(colony)
espectro <- cbind(y1,y2,y3,y4,y5)
dados <- data.frame(avaliacao = I(as.matrix(avaliacao)), bands = 
I(as.matrix(espectro)))


## Create a design matrix of the contrasts for "avaliacao"
comb <- cbind(combn(unique(c(dados$avaliacao,dados$avaliacao)), 
2)[2:1,])## Combinação entre os níveis
comb
coff.sig<-sapply(1:ncol(comb), function(x) with(dados,
       2*( (dados$avaliacao==comb[1,x] & dados$avaliacao == comb[2,x]) | 
(dados$avaliacao==comb[1,x] & dados$avaliacao == comb[2,x])) -
        (dados$avaliacao == dados$avaliacao & dados$avaliacao %in% 
comb[,x]) ) )
coff<- sapply(1:ncol(comb), function(x) with(dados,
       2*( (dados$avaliacao==comb[1,x] & dados$avaliacao == comb[2,x]) | 
(dados$avaliacao==comb[1,x] & dados$avaliacao == comb[2,x])) -
        (dados$avaliacao == dados$avaliacao & dados$avaliacao %in% 
comb[,x]) ) )


coff## Problem here!!!!!!!!!!


## Create a factor for each of the orthogonal "avaliacao" contrasts
col1vs2 <-  coff[, 1]
col1vs3 <-  coff[, 2]
col2vs3 <-  coff[, 3]

## Do the pm-manova with the full model
man.col<-adonis(bands ~ avaliacao, data=dados,permutations=99)
man.col

## Do the pm-manova with the orthogonal contrasts for "avaliacao"
## and the interaction contrasts of interest
man.col2 <- adonis(dados$bands ~ col1vs2 + col1vs3 + col2vs3, 
permutations=99)
man.col2
#
#
#Better coefficients
#
man.col2$coefficients[3,abs(man.col2$coefficients[3,])>1]
#
#<END>--------------------------------------------------------------------------------------------------

-- 
======================================================================
Alexandre dos Santos
Forest Protection
Federal Institute of Education, Science and Tecnology - IFMT
Campus Cáceres
Caixa Postal 244
Avenida dos Ramires, s/n
Bairro: Distrito Industrial
Cáceres - MT - BRAZIL            ZIP-CODE: 78.200-000
Tel: (+55) 65 8132-8112 (TIM)   (+55) 65 9686-6970 (VIVO)

         alexandre.santos at cas.ifmt.edu.br
Lattes: http://lattes.cnpq.br/1360403201088680
======================================================================


---
Este email foi escaneado pelo Avast antivírus.
http://www.avast.com