[vsni.co.uk] Contact us
Author Message
Post new topic

  ASReml  ~  Implementing spatial matrixes ASREML standalone (problems)

fra
Posted: Fri Apr 06, 2018 3:24 pm Reply with quote
Joined: 05 Apr 2018 Posts: 1
Hi all,

I am trying to implement spatial matrixes of pairwise individual distances in asreml. This should work in the same way as a normal genetic relatedness matrix, to account for space sharing between individual. I am not including the pedigree, just this spatial matrix.

I am following a similar approach as Regan, C. E., et al in. "Accounting for female space sharing in St. Kilda Soay sheep (Ovis aries) results in little change in heritability estimates." Journal of evolutionary biology 30.1 (2017): 96-111. They provide code for ASREML-R but I would like to use the standalone version.

I prepared my spatial matrix as a dense lower triangle matrix but I am having troubles with the model, as I get "Singularity in Average Information Matrix" errors. Could anyone spot where I am making a mistake?

thanks a lot!


Here are the relevant bits of code:

M95.grm !DENSEGRM !NSD
#this reads my lower triangle (incl. diagonal) matrix of spatial distance. The matrix has the same number of rows as my dataset

data5 for asreml.txt !SKIP 1 !STEP 0.001 !MAXIT 500 !DOPART 6 !display 2 !FCON !DDF 1 !MVINCLUDE !SLOW 5 !EXTRA 5 !CONTINUE !AISING
#this reads my datafile

!part 6
foc_tot_cross_SC ~ mu !r giv(id_season_year,1)
#I have a phenotype as y variable, and I want to include my spatial matrix with 'giv'. My random effect is a unique row identifier for each individual each season.


This is my output:

Notice: 192 singularities detected in design matrix.
1 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
Notice: 1 singularities appeared in Average Information matrix
This could be a problem of scale or a problem with the model.
It is preferable to revise the model to remove the singularity.
Specify !AISING qualifier to force the job to continue.
2 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
3 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
4 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
5 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
6 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
7 LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
Final parameter values 0.1544E+13

- - - Results from analysis of foc_tot_cross_SC - - -
Akaike Information Criterion 5885.73 (assuming 2 parameters).
Bayesian Information Criterion 5893.69

Model_Term Gamma Sigma Sigma/SE % C
giv(id_season_year,1) GRM_V 586 0.154408E+13 4.28278 0.00 0 S
Residual SCA_V 395 1.00000 0.277368E-11 0.00 0 ?
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
Notice: The DenDF values are calculated ignoring fixed/boundary/singular
variance parameters using numerical derivatives.

Solution Standard Error T-value T-prev
56 giv(id_season_year,1) 395 effects fitted (+ 191 singular)
Finished: 06 Apr 2017 11:00:19.773 Singularity in Average Information Matrix
View user's profile Send private message
Arthur
Posted: Sat Apr 07, 2018 9:43 pm Reply with quote
Joined: 05 Aug 2008 Posts: 471 Location: Orange, NSW
Dear FRA,

It is not obvious what the issue is.

It could be related to the scale of the response variable,
It is most likely the GRM is not inverting properly,


LogL=-2940.87 S2= 0.27737E-11 395 df 0.1544E+13
the initial value of the variance component seems extraordinarly high (0.1544E+13)
and the residual extraordinarily low (0.27737E-11).

If you email the whole job to arthur@vsni.co.uk
I will have a look and see what's going wrong.

If you can't do that, at least send the whole .as file and the .asr file.

_________________
Arthur Gilmour

Retired Principal Research Scientist (Biometrics)
View user's profile Send private message Send e-mail Visit poster's website

Display posts from previous:  

All times are GMT
Page 1 of 1
Post new topic

Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum