Online Program

Friday, February 21
PS2 Poster Session II & Refreshments Fri, Feb 21, 4:45 PM - 6:15 PM
Bayshore II-IV

Mixed Effects Model for Comparing Treatments That Alter Length of Life in the C. elegans Model (302800)

Robbie Beyl, Pennington Biomedical Research Center 
*Jeffrey H. Burton, Pennington Biomedical Research Center 
William D. Johnson, Pennington Biomedical Research Center 

Keywords: Mixed effects model, group-level data, C. elegans, survival

In some laboratory experiments, researchers can only measure survival on group-level data. One such study concerns survival of C. elegans (roundworms) following treatment with the insecticide imidacloprid. Groups of worms were randomized to wells (replicates) containing one of three doses of insecticide, and counts of living worms within each well were recorded at multiple time points. Due to worm reproduction, counts of living worms increased within some wells from one observation to the next. Since it is not possible to track individual worms, one cannot measure individual survival times. Thus, traditional survival analysis is not appropriate for these data. Instead, a generalized linear mixed model assuming a Poisson distributed response (counts) may be used. Here, a Poisson model is used to obtain estimates of least squares means of counts of living worms at each time. From the means, proportions are calculated, used to estimate survival probabilities for each dose, and compared between doses at the final time point. The purpose is to show that this method is a viable alternative to traditional survival analysis when data are available only on the group level.