Re: Singularity in Average Information Matrix

Dear all,
When I fit qtl effect in this data, I get a singular information
matrix. I have looked at the data and I do not seem to see anything
obvious that I am missing. This happens even when I remove all the fixed
effects. Please see below, part of .as file and .asr file:

Part of As File:

  nqtl 814 !I
pd.prn
inv5.giv
no3fldat1.txt !MAXIT 100 !EXTRA 50

# Model 1 (Animal)
lnemat1 ~ mu !r animal nqtl

0 0 2
animal 1
animal 0 AINV 0.1 !GP

nqtl 1
nqtl 0 GIV1 0.1 !GP

ASR FILE:
Folder:
M:\data2007\GenoResults07\Suffolk\nematode\cleandata\Asreml\chr14\nemat
   animal !P
   sire !P
   dam !P
   flock !I
   year !I
   sex !I
   tob !I
   tor !I
   mtor !I
   dage !I
   mdage !I
   w8w !M0
   aos !M0
   scanwt !M0
   mscd !M0
   fatd !M0
   nqtl 814 !I
 A-inverse retrieved from ainverse.bin
 PEDIGREE [pd.prn ] has 814 identities, 2091 Non zero elements
  Reading inv5.giv skipping 0 header lines
 Inverse G structure of 814 rows having 112292 non zero cells read
from inv5.giv

 QUALIFIERS: !MAXIT 100 !EXTRA 50
 Reading no3fldat1.txt FREE FORMAT skipping 0 lines

 Univariate analysis of lnemat1
 Using 305 records of 305 read

  Model term Size #miss #zero MinNon0 Mean
MaxNon0
   1 animal !P 814 0 0 15.00 375.4
814.0
   2 sire !P 814 0 0 3.000 134.5
421.0
   3 dam !P 814 0 0 14.00 368.1
813.0
   4 flock 6 0 0 1 3.3049
6
   5 year 2 0 0 1 1.5115
2
   6 sex 2 0 0 1 1.4557
2
   7 tob 3 0 0 1 1.9475
3
   8 tor 3 0 0 1 1.7311
3
   9 mtor 2 0 0 1 1.6590
2
  10 dage 4 0 0 1 2.2066
4
  11 mdage 3 0 0 1 1.8000
3
  12 w8w 5 0 10.20 25.71
42.30
  13 aos 31 0 108.0 136.1
211.0
  14 scanwt 31 0 30.00 52.75
84.00
  15 mscd 31 0 21.00 31.58
41.00
  16 fatd 0 0 0.3300 13.08
99.00
  17 lnemat1 Variate 0 213 3.434 1.258
6.217
  18 lnemat2 0 0 0.6931 1.748
6.219
  19 bnemat 0 213 1.000 0.3016
1.000
  Warning: Fewer levels found in nqtl than specified
  20 nqtl 814 0 0 1 153.0000
305
  21 mu 1
    814 Ainverse 0.1000
 Structure for animal has 814 levels defined
    814 inv5.giv 0.1000
 Structure for nqtl has 814 levels defined
 Forming 1629 equations: 1 dense.
 Initial updates will be shrunk by factor 0.100
 Notice: Algebraic ANOVA Denominator DF calculation is not available
         Numerical derivatives will be used.

 ** Notice **: GIV matrix not positive definite: Negative pivots 17
814

   1 LogL=-348.670 S2= 3.0093 304 df 1.000 0.1000
0.1000

 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.

 Source Model terms Gamma Component Comp/SE
% C
 Variance 305 304 1.00000 3.01101 0.00
0 S
 animal Ainverse 814 0.166668 0.501840 2.88
27 P
 nqtl inv5.giv 814 0.481359E-02 0.144938E-01 0.00
-93 ?
 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.

 Analysis of Variance NumDF DenDF F_inc
Prob
  21 mu 1 60.4 53.12
<.001
 Notice: The DenDF values are calculated ignoring
fixed/boundary/singular
             variance parameters using numerical derivatives.

                     Estimate Standard Error T-value T-prev
  21 mu
                    1 1.13752 0.163233 6.97
   1 animal 814 effects fitted ( 172
are zero)
  20 nqtl 814 effects fitted ( 4
are zero)
 SLOPES FOR LOG(ABS(RES)) on LOG(PV) for Section 1
   1.08
 Finished: 20 Sep 2007 10:23:51.680 Singularity in Average Information
Matrix

Regards,

Ozzie

Email: Oswald.Matika_at_bbsrc.ac.uk
Web: www.roslin.ac.uk <<http://www.roslin.ac.uk>>

Roslin Institute is a company limited by guarantee, registered in
Scotland (registered number SC157100) and a Scottish Charity (registered
number SC023592). Our registered office is at Roslin, Midlothian, EH25
9PS. VAT registration number 847380013.

The information contained in this e-mail (including any attachments) is
confidential and is intended for the use of the addressee only. The
opinions expressed within this e-mail (including any attachments) are
the opinions of the sender and do not necessarily constitute those of
Roslin Institute (Edinburgh) ("the Institute") unless specifically
stated by a sender who is duly authorised to do so on behalf of the
Institute
Received on Mon Sep 20 2007 - 10:36:47 EST