Help using Base SAS procedures

Repeated Measures in Glimmix

Reply
New Contributor
Posts: 2

Repeated Measures in Glimmix

I have a data set that I need help analyzing. I have 20 animals that were broken into a light and heavy weight block. Within each block, animals were assigned to 1 of 2 treatments. A muscle biopsy was taken at 14 day intervals and gene expression was measured. I need to analyze the data with repeated measures over time, but I do not know the correct place to put my random effect of block into the code. My data looks like this for one variable at 2 of my time points:

 

DATA;

INPUT ID DAY TRT BLOCK GENE;

219 0 0 2 0.201 
223 0 0 1 1.000 
287 0 0 1 0.543 
296 0 0 2 0.455 
654 0 0 2 0.657 
668 0 0 1 0.845 
681 0 0 2 0.243 
692 0 0 1 0.446 
723 0 0 1 0.980 
920 0 0 2 1.002 
256 0 8 1 0.908 
653 0 8 1 0.634 
664 0 8 2 0.417 
674 0 8 1 0.554 
709 0 8 1 2.139 
710 0 8 2 1.083 
715 0 8 2 0.697 
745 0 8 1 0.477 
924 0 8 2 0.612 
934 0 8 2 0.464 
219 14 0 2 0.564 
223 14 0 1 0.851 
287 14 0 1 0.603 
296 14 0 2 0.437 
654 14 0 2 0.314 
668 14 0 1 0.640 
681 14 0 2 0.876 
692 14 0 1 0.367 
723 14 0 1 0.918 
920 14 0 2 0.663 
256 14 8 1 1.105 
653 14 8 1 0.555 
664 14 8 2 0.360
674 14 8 1 0.622 
709 14 8 1 0.635 
710 14 8 2 0.554 
715 14 8 2 0.594 
745 14 8 1 0.896 
924 14 8 2 0.842 
934 14 8 2 0.530 

;

 and some of the code I have been shown looks as follows:

 

1)

PROC GLIMMIX;
Class TRT DAY BLOCK ID;
Model AMPK = TRT|DAY / ddfm = SATTERTH;
Random BLOCK / SUBJECT = ID(TRT) TYPE = CS;
LSMEANS TRT|DAY;
RUN;

 

2)

PROC GLIMMIX;
Class TRT DAY BLOCK ID;
Model AMPK = TRT|DAY / ddfm = SATTERTH;
Random _RESIDUAL_/SUBJECT = BLOCK(TRT) TYPE = CS;
LSMEANS TRT|DAY;
RUN;

 

3)

PROC GLIMMIX;
Class TRT DAY BLOCK ID;
Model AMPK = TRT|DAY / ddfm = SATTERTH;
Random BLOCK;
Random _residual_ / SUBJECT = ID(TRT);
LSMEANS TRT|DAY;
RUN;

 

I need to get at the treatment, day, and treatment*day effect, but since the study was designed with the random effct of block, I need to know the proper place to include this in my model. I am somewhat lost on the subject and any help would be appreciated.

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