Hello, I have been entertaining the notion that high Ct values, which are generally used as an indicator of no replication due to various reasons, might actually be a response worth prediction via a binomial regression. So for example, does a predictor variable (Sex, treatment, etc.) have a significant effect on the probability of a high or "undetermined" Ct. Is this sound reasoning? If so, does anyone know of an implementation to apply a binomial regression to high-throughput data? (In my case, it would be a miR HTqPCR array). Thanks for any discussion/suggestions, Chris -- Christopher Howerton, PhD Postdoctoral Researcher (215) 898-1368 University of Pennsylvania 201E Vet 3800 Spruce Street Philadelphia, PA 19104-6046