Help using Base SAS procedures

Repeated Measures using a Mixed Model ""Error: Stopped because of infinite likelihood""

Reply
Frequent Learner
Posts: 1

Repeated Measures using a Mixed Model ""Error: Stopped because of infinite likelihood""

Hello,  I’m testing for percent compositional changes in various bacteria by testing cecal/fecal bacteria at three time points. (Pre-Treatment, Post Antibiotics, and Final (Post AB + Donor cecal transplant). Some of my data have zero values due to the antibiotic knockdown.   In some cases a given bacterium was knocked down to zero in all rats.  In other cases I have zero amount of a bacterium before treatment (PreAB) and an a percent amount after treatment (PostAB).   I unfortunately keep recieving the error: "WARNING: Stopped because of infinite likelihood".  Do I rerun without certain time points and note in text or work with percent differences across time periods?  Any help would be greatly appreciated.  Thanks!

 

 

 

Code:

dm'log; clear; output; clear';
options nodate nocenter pageno=1 Ls=78 ps=53;
title 'ZDF Transplant Repeated Measures Tenericutes';
ods rtf file='\\tsclient\C\ZDFTrans\Tenericutes.rtf';
*ods html file='C:/temp/ZDFTrans/Urine.html';

*********************************************************************************;
***   The following data are measurements of bacteria over the following 3    ***;
*** timepoints (TP) (Pre-Antibiotics, Post-Antibiotics and a cecal transplant)***;
***   in lean ZDF rats.  Cecal contents donated by either(LeanRS or LeanHF)   ***;    
*********************************************************************************;

**********************************************************************************;
*** Measurements were collected: PreAB(day 1) PostAB(day 4) Transplant (day 45)***;
**********************************************************************************;
*Rat = Order # (1-4, 5-8) [4]
 Group = Diet/Transplant type [2]
 TP = TimePoint [3]
 y = Amount of an observed bacteria at a given timpoint (ex. Bacteroidetes)
 4x2x3 = 24 datalines
 ;

data ZDFTrans;

     input Rat Group $ y1 y2 y3;

     TP=1; y=y1;  output;

     TP=2; y=y2;  output;

     TP=3; y=y3;  output;

     drop y1-y3;

      length y 8;

      label y = 'Mogibacteriaceae';

     datalines;

1    LeanRS    0.016279917    0.045808967    0.000442413
2    LeanRS    0            0.141046474    0
3    LeanRS    0.000172265    0.04442344    0.000229148
4    LeanRS    0            0.03313253    0.000247709
5    LeanHF    0            0.280766852    0.000225581
6    LeanHF    0.000220751    0.053085143    0.000451569
7    LeanHF    0            0.049773756    0.000547645
8    LeanHF    0.000508001    0.022063512    0.000213129

 

 

 

 

 

   proc mixed data=ZDFTrans method=ml covtest;

      class Rat TP Group;

      model y = Group TP Group*TP / outp=resids s;

        random Rat(Group);

        repeated / type=un subject=Rat r;

        lsmeans Group|TP / pdiff cl;

        *ods listing exclude diffs lsmeans;

run;

 

   proc univariate data=resids plot normal;

   var resid;

   run;

 

ods rtf close;

ods html close;

run;

quit;

Ask a Question
Discussion stats
  • 0 replies
  • 156 views
  • 0 likes
  • 1 in conversation