R help - Jan 2008 - Error in solve.default message using lme

Dear All,

 

I?m using lme to analyze a time series gene expression data. I don?t have
all animals in all times, the number of animals in each time is different
and I have lots of NA values.

An example file can be downloaded at lbmp.fcav.unesp.br/leonardo and the
code used is the following:

 

teste<-read.table(file='example.csv',sep=',',header=T)

B<-1366

library(nlme)

pValueTime<-matrix(nrow=B,ncol=2)

colnames(pValueTime)<-c('geneID','pValue.Time')

 

for (i in 1:B) {

                print(i)

                gene = teste$gene [i]

                keep<-teste$gene==gene

                MTemp =as.numeric( teste$M [keep])

                timeTemp = as.factor(teste$time [keep])

                animalTemp = as.factor(teste$animal [keep])

 

  tempMM.HA<-lme(MTemp~timeTemp,random=~1|animalTemp,method='ML',
na.action=na.omit)

  tempMM.H0<-lme(MTemp~1,random=~1|animalTemp,method='ML',
na.action=na.omit)

   pValueTime[i,1]<-gene

   pValueTime[i,2]<-anova(tempMM.HA,tempMM.H0)[2,9]

}

 

I?m getting the following error message:

 

Error in solve.default(estimates[dimE[1] - (p:1), dimE[2] - (p:1), drop FALSE])
:

  system is computationally singular: reciprocal condition number 1.64483e-17

 

I don?t know what this message means, so any help/suggestion will be very
appreciated.

 

sessionInfo()

R version 2.6.0 (2007-10-03) 

i386-pc-mingw32 

 

locale:

LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
States.1252;LC_MONETARY=English_United
States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252

 

attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base     

 

other attached packages:

[1] nlme_3.1-86

 

loaded via a namespace (and not attached):

[1] grid_2.6.0     lattice_0.17-2 tools_2.6.0   

 

 

Best regards,

 

Leonardo

 

____________________________________

MSc. Leonardo Bernardes da Rocha

Doctorate Student - Animal Science

Department of Tecnology

College of Agricultural and Veterinary Sciences

S?o Paulo State University

Jaboticabal, SP - Brazil

Phone: +55 16 32092675 (Ext. 206)

Email: leonardo.rocha at posgrad.fcav.unesp.br

____________________________________



-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: example.txt
Url:
https://stat.ethz.ch/pipermail/r-help/attachments/20080114/8ea71c80/attachment-0002.txt

What it probably means is that you can't make a silk purse out of a
sow's
ear: all those missings in your your data do not allow you to estimate the
model(s) that you've specified. 

If you don't know what this means, I suggest you seek the help of a local
statistician.

-- 
Bert Gunter
Genentech


-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On
Behalf Of Leonardo
Sent: Monday, January 14, 2008 8:51 AM
To: r-help at r-project.org
Subject: [R] Error in solve.default message using lme

Dear All,

 

I?m using lme to analyze a time series gene expression data. I don?t have
all animals in all times, the number of animals in each time is different
and I have lots of NA values.

An example file can be downloaded at lbmp.fcav.unesp.br/leonardo and the
code used is the following:

 

teste<-read.table(file='example.csv',sep=',',header=T)

B<-1366

library(nlme)

pValueTime<-matrix(nrow=B,ncol=2)

colnames(pValueTime)<-c('geneID','pValue.Time')

 

for (i in 1:B) {

                print(i)

                gene = teste$gene [i]

                keep<-teste$gene==gene

                MTemp =as.numeric( teste$M [keep])

                timeTemp = as.factor(teste$time [keep])

                animalTemp = as.factor(teste$animal [keep])

 

  tempMM.HA<-lme(MTemp~timeTemp,random=~1|animalTemp,method='ML',
na.action=na.omit)

  tempMM.H0<-lme(MTemp~1,random=~1|animalTemp,method='ML',
na.action=na.omit)

   pValueTime[i,1]<-gene

   pValueTime[i,2]<-anova(tempMM.HA,tempMM.H0)[2,9]

}

 

I?m getting the following error message:

 

Error in solve.default(estimates[dimE[1] - (p:1), dimE[2] - (p:1), drop FALSE])
:

  system is computationally singular: reciprocal condition number 1.64483e-17

 

I don?t know what this message means, so any help/suggestion will be very
appreciated.

 

sessionInfo()

R version 2.6.0 (2007-10-03) 

i386-pc-mingw32 

 

locale:

LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
States.1252;LC_MONETARY=English_United
States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252

 

attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base     

 

other attached packages:

[1] nlme_3.1-86

 

loaded via a namespace (and not attached):

[1] grid_2.6.0     lattice_0.17-2 tools_2.6.0   

 

 

Best regards,

 

Leonardo

 

____________________________________

MSc. Leonardo Bernardes da Rocha

Doctorate Student - Animal Science

Department of Tecnology

College of Agricultural and Veterinary Sciences

S?o Paulo State University

Jaboticabal, SP - Brazil

Phone: +55 16 32092675 (Ext. 206)

Email: leonardo.rocha at posgrad.fcav.unesp.br