Re: warning message

dear derek

I have discussed this with David Butler.

His take on this is
"
the desc on p70 is for the 'raw' object and the scaling is done in
summary.asreml(). so, if you had a negative component you would presumably
get some -ve numbers in object$vcoeff which would cause R to complain
about sqrt(-ve) numbers when summary.asreml() was run.

this happens in the bloodworm example in the manual though summary() is
not included in the illustration.
"
I tried the same on an object and got sensible results ie non NANs

eg

> > summary(fababeans.asr)$coef.ra[1:12,]
> solution std error z ratio
> Genotype_ 0.07963427 0.05533198 1.4392086
> Genotype_ 0.15675743 0.14728191 1.0643359
> Genotype_ -0.15892526 0.09003410 -1.7651673
> Genotype_ -0.22162121 0.09046157 -2.4498934
> Genotype_ -0.21007788 0.07972240 -2.6351173
> Genotype_ -0.07384444 0.07992000 -0.9239795
> Genotype_ -0.12358285 0.09961888 -1.2405565
> Genotype_ -0.12970456 0.12070645 -1.0745455
> Genotype_ 0.17536016 0.09926751 1.7665414
> Genotype_ -0.28158856 0.12548071 -2.2440785
> Genotype_ 0.17219450 0.06240132 2.7594688
> Genotype_ 0.12521249 0.09113118 1.3739808
>
> > sqrt(fababeans.asr$vcoeff$random[1:12])
> [1] 0.05533198 0.14728191 0.09003410 0.09046157 0.07972240 0.07992000
> [7] 0.09961888 0.12070645 0.09926751 0.12548071 0.06240132 0.09113118
> >

hope this helps

brian

Brian Cullis
Research Leader, Biometrics &
Principal Research Scientist
NSW Department of Primary Industries
Wagga Wagga Agricultural Institute

Professor,
Faculty of Agriculture, Food & Natural Resources
The University of Sydney

Phone: 61 2 6938 1855
Fax: 61 2 6938 1809
Mobile: 0439 448 591

Derek Roff <derekr_at_UCR.EDU>
Sent by: ASReml users discussion group <ASREML-L_at_AGRIC.NSW.GOV.AU>
06/11/2007 06:25 AM
Please respond to
ASReml users discussion group <ASREML-L_at_AGRIC.NSW.GOV.AU>

To
ASREML-L_at_AGRIC.NSW.GOV.AU
cc

Subject
warning message

Hi,
In running R for a two trait model I get the following
warning message
Warning message:
In sqrt(s2 * object$vcoeff$random) : NaNs produced

The heritability estimate is clearly wrong, as indicated by
the program run for a single trait, which does not give this
message.

Can anyone tell me what the message refers to and how I
might resolve the problem?
Thanks
Derek
Derek Roff,
Department of Biology,
University of California,
Riverside,
CA 92521
USA
Received on Tue Nov 09 2007 - 08:51:18 EST