University of Arizona
My data is collected based upon animal physiology, but I am only evaluating the glucocorticoid metabolite concentrations (feces) so I do not conduct any testing on animals, but I did collect their feces in 2015, 2016, in collaboration with the Phoenix Zoo’s conservation center under their research guidelines.
My work is to track the physiology of Mt. Graham red squirrels longitudinally to determine if when they are in estrus. This research is part of the Koprowski lab, and John Koprowski is my advisor. I am working to develop an ex situ breeding management strategy for this endangered species. Right now I am attempting to confirm the existence of polyestrus, and to determine the periodicy of their estrus cycle. I am also trying to confirm whether these cycles occur at around the same time annually. I have attached a dataset from one female that I am would like to determine the best way to statistically confirm or rule out periodicy, and repeatability of the cycles.
Client: Stuart Wells (University of Arizona – SNRE
Consultants: Elmira, Haozhe, Samir, Amy (author)
15 October 2019, 3-4:30pm
A. Overview of Client’s Problem
Stuart is evaluating the glucocorticoid metabolite concentrations (feces) to track the physiology of Mt. Graham red squirrels longitudinally to determine when they are in estrus. Based on the literature, these red squirrels are known to be receptive 1 day a year for 6 hours. However, Stuart observed some periodicity of their estrus cycle according to physiological data and wants to confirm the existence of polyestrous. Furthermore, he wants to confirm that these cycles occur at around the same time annually.
1) Data details:
The data includes the physiology data for a single female squirrel over a similar time frame in 2015 and 2016. The dataset has progesterone, cortisol, and estradiol levels measured from April to August.
2) Background information:
It is known that these female red squirrels are receptive only 1 day per year for 6 hours, but multiple peaks of estradiol levels were observed based on the physiological data that Stuart obtained in the lab. Since there are multiple peaks, he wants to know which of these peaks is significantly different, and if the peak is really a peak. However, there is no baseline for the data, so it was not possible for him to identify peaks and to perform follow-up analyses according to the peak identification. He believed that some of these multiple peaks are really peaks because not only the estradiol level, but the progesterone level has a peak shortly after a peak occurs in the estradiol data. When both progesterone and estradiol have peaks, this indicates the estrous cycle. Furthermore, this pattern in progesterone data suggests that the red squirrels go through spontaneous ovulation rather than induced ovulation. Also, there are other factors that can affect the estradiol level, such as temperature, behavior features (aggression), so he also provided the cortisol level data.
3) Outcome of interest:
- Baseline function to identify a peak
- Statistical test to compare distributions to determine if peaks occur at similar times
- Statistical test of parameters: distance between two peaks and the height of each peaks
IV. Next steps
1) Our suggestions
- First, we need to figure out ways to determine the baseline that makes sense biologically to carry on other statistical tests and analyses
- One method can be setting the minimum value from the first year and use this value for the rest of years as baseline
2) Things to be discussed in class
- How can we determine the baseline?