Re: Asreml_R help: Is the model wrong?

Dear Li,

I fitted a model in Standalone ASReml 3 and that returned DF.
My model is a little different from yours and shows trial.treatment.date interaction.

The ASReml code was

Title: ecoli_633.
 trial !A
 treatment !A
 sdate
 block !A
 chamber *
 plate *
 logcount
 
ecoli_633.txt !SKIP 1 !WMF !FCON
!SPLINE spl(sdate,8) 14 28 56 75 100 125 150 180
logcount ~ mu trial*treat*sdate , # Specify fixed model
      !r spl(sdate,8) trial.spl(sdate,8) trial.block/chamber fac(sdate) treat.spl(sdate,8) # Specify random model

predict trial treat sdate 7 14 20 28 34 48 69 83 100 112 125 133 150 174 205 !PLOT ^super treat ^xaxis sdate !ignore fac(sdate)

  2 LogL=-556.754 S2= 0.79467 1146 df
   3 LogL=-556.754 S2= 0.79467 1146 df
   4 LogL=-556.754 S2= 0.79467 1146 df

          - - - Results from analysis of logcount - - -

          Approximate stratum variance decomposition
 Stratum Degrees-Freedom Variance Component Coefficients
 treat.spl(sdate,8 8.37 2.11648 7.8 0.0 0.1 0.0 1.0
 trial.spl(sdate,8 13.59 9.93321 0.0 5.3 4.8 -0.3 1.0
 fac(sdate) 7.24 34.5840 0.0 0.0 89.4 1.4 1.0
 trial.block.chamb 17.49 2.06647 0.0 0.0 0.0 58.6 1.0
 Residual Variance 1099.31 0.794673 0.0 0.0 0.0 0.0 1.0

 Source Model terms Gamma Component Comp/SE % C
 spl(sdate,8) 6 6 0.319382E-09 0.253804E-09 0.00 0 B
 trial.block 9 9 0.404772E-09 0.321661E-09 0.00 0 B
 treat.spl(sdate,8) 12 12 0.200197 0.159091 1.20 0 P
 trial.spl(sdate,8) 18 18 1.74044 1.38308 1.86 0 P
 fac(sdate) 21 21 0.475203 0.377631 1.86 0 P
 trial.block.chamber 27 27 0.272976E-01 0.216926E-01 1.82 0 P
 Variance 1158 1146 1.00000 0.794673 23.44 0 P
 Warning: Code B - fixed at a boundary (!GP) F - fixed by user
               ? - liable to change from P to B P - positive definite
               C - Constrained by user (!VCC) U - unbounded
               S - Singular Information matrix
 S means there is no information in the data for this parameter.
 Very small components with Comp/SE ratios of zero sometimes indicate poor
           scaling. Consider rescaling the design matrix in such cases.

                                   Wald F statistics
     Source of Variation NumDF DenDF_con F-inc F-con M P-con
   9 mu 1 8.0 282.69 194.44 . <.001
   1 trial 2 27.8 12.09 13.60 A <.001
   2 treatment 1 1107.1 76.58 76.81 A <.001
   3 sdate 1 8.0 33.98 33.33 A <.001

  10 trial.treat 2 1107.3 263.64 250.14 B <.001
  11 trial.sdate 2 13.3 0.86 0.88 B 0.440
  12 treat.sdate 1 1104.4 12.52 12.52 B <.001
  13 trial.treat.sdate 2 1103.0 13.79 13.79 C <.001
 

Note the use of fac(sdate) to smooth the predicted response profile
The pvs file is attached.

------------------------
Arthur Gilmour

Retired Principal Research Scientist (Biometrics)

-------------------- m2f --------------------

Sent using Mail2Forum (http://www.mail2forum.com).

Read this topic online here:
http://www.vsni.co.uk/forum/viewtopic.php?p=804#804

-------------------- m2f --------------------

Attachments:
http://www.vsni.co.uk/forum/files/ecoli_pvs_612.txt
Received on Fri Sep 03 2009 - 13:06:57 EST