Re: error structure in bivariate models with ASReml3

Dear Arthur,

Thank you for your swift answer. The fact is, I have really 3742 records, and I have not used !ASUV, but I guess ASReml was counting 7364 records when amounting all records where LD was available, and all where AD was available. Your option unfortunately gave the same error message. However, since you pointed out that this was probably because of missing records, I restricted the dataset to the 3130 records where both LD, AD and SEX are known, and the model runs correctly, and converges. Below is the full output from the .asd.
Just a couple of extra questions if you have the time, to make sure I am in the right track:
- “6260 records assumed pre-sorted 2 within 3130” : this means that the two traits in the bivariate model have both 3130 values recorded, is that right?
- it seems I no longer need the !S2==1 option which I still use in my univariate models, else I get an extra error variance. Why do I not need it now?

Thank you so much for your help,
Anne
--
Reading Breeders_LD_AD_restricted.asd FREE FORMAT skipping 1 lines

 Bivariate analysis of AD and LD
 Using 3130 records of 3130 read

  Model term Size #miss #zero MinNon0 Mean MaxNon0
   1 ANIMAL !P 4023 0 0 2.000 1224. 3784.
   2 SEX 2 0 0 1 1.5054 2
   3 YEAR 22 214 0 1 10.3735 22
   4 STATUS 15 0 0 1 7.1827 15
   5 CS 0 0 1.000 1.000 1.000
   6 AGE 0 0 1.000 2.518 3.000
   7 AGE2 214 0 2.000 6.535 22.00
   8 NBFLEDG 214 0 4.000 51.67 484.0
   9 LD 3 0 1686 1 0.6665 3
  10 WEIGHT 0 0 121.0 143.3 280.0
  11 AD 161 1373 0 100 131.0852 161
  12 REPRO 66 0 0 1 19.2514 66
  13 MO 2 0 0 1 1.8073 2
  14 FA 191 1652 0 1 60.3281 191
  15 Year1Breed 221 1652 0 1 77.2463 221
  16 Brood 17 389 0 1 7.9321 17
  17 Trait 2
  18 Trait.SEX 4 17 Trait : 2 2 SEX : 2
   3130 identity
      2 UnStructure 19.3444 0.1331 0.1984
    6260 records assumed pre-sorted 2 within 3130
 Forming 6 equations: 6 dense.
 Initial updates will be shrunk by factor 0.316
 Notice: Algebraic ANOVA Denominator DF calculation is not available
         Numerical derivatives will be used.
 Notice: 2 singularities detected in design matrix.
   1 LogL=-8768.97 S2= 1.0000 4883 df
   2 LogL=-7659.44 S2= 1.0000 4883 df
   3 LogL=-6519.16 S2= 1.0000 4883 df
   4 LogL=-5766.35 S2= 1.0000 4883 df
   5 LogL=-5628.51 S2= 1.0000 4883 df
   6 LogL=-5620.30 S2= 1.0000 4883 df
   7 LogL=-5620.26 S2= 1.0000 4883 df
   8 LogL=-5620.26 S2= 1.0000 4883 df

 Source Model terms Gamma Component Comp/SE % C
 Residual UnStructured 1 1 68.6195 68.6195 29.63 0 U
 Residual UnStructured 2 1 0.622668 0.622668 3.82 0 U
 Residual UnStructured 2 2 0.708517 0.708517 39.55 0 U
 Covariance/Variance/Correlation Matrix UnStructured Residual
   68.62 0.8930E-01
  0.6227 0.7085

 Analysis of Variance NumDF DenDF F_inc Prob
  17 Trait 2 2437.1220641.65 <.001
  18 Trait.SEX 2 2439.2 7.75 <.001
 Notice: The DenDF values are calculated ignoring fixed/boundary/singular
             variance parameters using numerical derivatives.

                     Estimate Standard Error T-value T-prev
  18 Trait.SEX
                    2 -1.40235 0.395489 -3.55
                    4 0.443123E-01 0.300925E-01 1.47
  17 Trait
                    1 131.825 0.289087 456.00
                    2 0.644057 0.213939E-01 30.10 -454.72
 Finished: 22 Dec 2009 16:14:19.119 LogL Converged


------------------------------------
Dr Anne Charmantier
  CEFE-CNRS, UMR 5175
  1919, route de Mende
  F34293 Montpellier Cedex 5
  France

http://annecharmantier.free.fr

 Tel : +33 4 67 61 32 05
 Fax : +33 4 67 41 21 38
 Email: anne.charmantier_at_cefe.cnrs.fr
-------------------------------------

-----Message d'origine-----
De : ASReml users discussion group [mailto:ASREML-L_at_DPI.NSW.GOV.AU] De la part de Arthur Gilmour
Envoyé : 22 December 2009 10:32
À : ASREML-L_at_DPI.NSW.GOV.AU
Objet : Re: error structure in bivariate models with ASReml3

Dear Anne.



First, let me advise everyone reading this that I will not be reading my

email after 28 Dec until after 5 Feb.



Now,

The problem will be related to !MVREMOVE although I suspect you have not

given us the full story because I can't see how a problem arises

with the model you have specified.



ASReml 1 did less verification, and so is likely just not to have

detected the problem.



You have 3742 records, and you say 3130 have LD, AD and SEX.



But ASReml is using 7365 (not 6260) records so maybe you are

using !ASUV (which is why you would also need !S2==1)

and ASReml is therefor ignoring records with SEX unknown,

and individual observations missing.



You may be able to wrk out exactly what is happening by looking at the

data summary, noting where the missing values are.





Anyway, I suggest you try a job like



Bivariate with missing

  SEX * !Dv *

  LD !Dv *

  AD !Dv *

LDAD.dat



AD LD ~ Trait Trait.SEX

1 2 0

0

Trait 0 US

3 * 0



!Dv * should discard the records with missing values in the particular

trait.



This job should then report that it is using 3130 records (in the data

summary) and analyse those records.



Best wishes for Christmas and the New Year.





On Tue, 2009-12-22 at 09:50 +0100, Anne CHARMANTIER wrote:

> Dear Brian and Arthur,

>

>

>

> I am running into problems trying to run simple bivariate models and

> wonder whether it is because it is the first time I am using ASReml3.

> This is a study of 3742 records of laying date (LD) and arrival date

> from migration (AD). We have some missing records, but 3130 records

> with LD, AD and sex known (we are using MVREMOVE to restrict to these

> records).

>

>

>

> Here is the simple first model we are trying to run.

>

> - With ASReml1, I would have used the following syntax:

>

>

>

> #############################################################

>

>

>

> AD LD ~ Trait Trait.SEX

>

>

>

> 1 2 0

>

> 0

>

> Trait 0 US

>

> 3 * 0

>

>

>

> - Using the new ASReml3, we have used your user guide to rewrite the

> code as (although the guide says the previous is equivalent):

>

> AD LD ~ Trait Trait.SEX

>

>

>

> 1 2 0

>

> 0 0 ID

>

> Trait 0 US

>

> 3 * 0

>

>

>

> Running both these models with ASReml2, we get the conclusion error

> message of “structures are wrong size”. The intermediate message is:

>

> Error structures imply 7364 data cells but the data has 7365

>

> Fault 1 Error structures are wrong size

>

> Running them with ASReml3, we get an error message with unfinished

> sentence:

>

> “Warning: Dropped records were not evenly distributed across the”.

>

> I have a feeling this could be linked to the redundancy in residual

> variances that appeared with ASReml2, that we dealt with using !

> S2==1, but here iterations do not even start..

>

>

>

> Best wishes for Christmas and the New Year to all ASReml users,

>

> Anne

>

>

>

>

>

>

> This message is intended for the addressee named and may contain confidential information. If you are not the intended recipient, please delete it and notify the sender. Views expressed in this message are those of the individual sender, and are not necessarily the views of their organisation.

>

> --

> passerelle antivirus du campus CNRS de Montpellier

> --

> --

> passerelle antivirus du campus CNRS de Montpellier

> --

>

>

>

> [한메일]이메일 청구서함 필요지수

> 는?

>

>

> 미아찾기 캠페인

--

Best Wishes,

Arthur Gilmour



Adjunct Professor

School of Computing and Mathematics

Charles Sturt University



Joseph inherited the title 'King of the Jews', being the leading

descendent of King David. However, neither the Romans or King Herod

were about to recognize him as such. After he contracted to marry

Mary, he discovered she was pregnant. He was surprised and

disappointed, and contemplated a quiet divorce; to proceed would

confer his title on her son. But an angel appeared to him in a

dream saying, "Joseph, Son of David, do not be afraid to take to

you Mary, your wife, for that which is conceived in her is of the

Holy Spirit. She will have a son whom you shall name JESUS, for He

shall save His people from their sins."



When Jesus was born, Herod killed all the babies of Bethlehem under

two years but Joseph saved Jesus, taking Him to Egypt until Herod

died.



Ultimately, the Romans crucified Him with the charge 'Jesus of

Nazareth, the King of the Jews'. But He rose from the dead the

following Saturday night and is now waiting in Heaven for the time

God has set for Him to return and take up His title and throne in

Jerusalem.

 

Mobile Number +61 427 227 468

Home phone +61 2 6364 3288 Skype: Arthur.Gilmour



http://www.CargoVale.com.au/asreml



Travel:



Touring S E Asia 29Dec - 6 Feb

WCGALP 2010 Aug 1-6



This message is intended for the addressee named and may contain confidential information. If you are not the intended recipient, please delete it and notify the sender. Views expressed in this message are those of the individual sender, and are not necessarily the views of their organisation.



--
passerelle antivirus du campus CNRS de Montpellier
--

-- 
passerelle antivirus du campus CNRS de Montpellier
--