Our World Statistics Day conversations have been a great reminder of how much statistics can inform our lives. Do you have an example of how statistics has made a difference in your life? Share your story with the Community!
Choose Language Hide Translation Bar
Level V

## Reef coral data from the Austral Islands (French Polynesia) and the Cook Islands

This JMP worksheet contains environmental data from the Austral Islands of French Polynesia and the Cook Islands, as well as molecular+physiological data from 43 coral samples (one of which has been excluded since it was a non-target species). My goals were to look at the relationship between environment and physiology, as well as to identify outliers. Although I measured 10 response variables in the coral samples (maximum colony length, planar surface area, RNA/DNA ratio, Symbiodinium GCP [density of dinoflagellates within the coral], and six gene mRNAs), I only considered RNA/DNA ratio, Symbiodinium GCP, and the first five genes in the data analysis. I used both univariate approaches (e.g., ANOVA and stepwise regression) and multivariate approaches (MANOVA, PCA, MDS) and uncovered that, in general, the host coral species was the most important "environmental" factor (i.e., the most important driver of variation). For multivariate approaches, I converted the values to z-scores such that they would all be on the same scale. I calculated Mahalanobis distance for each sample, as I actually think this is the most interesting parameter to measure in a coral (in which we do not KNOW what constitutes healthy, as we do in humans); the degree of deviation of a sample from the mean centroid may, as it turn out, be the most important factor for determining whether a coral is susceptible to climate change. See the PCA plot written to the table, for instance. I think you would agree that colony 88.2 might be "interesting." Most people would throw it out, though, due to its outlier status! Now the question remains: are outliers the "strongest" corals or the weakest? I don't know, but aim to find out.

In short, I am trying to develop a way to identify aberrantly behaving corals in the absence of knowing what constitutes healthy. This is a tricky question, and I don't know if anyone has dealt with this issue in other fields (I know for sure no one has in the coral reef field.). If you have any ideas about this issue, please do not hesitate to chime in.

https://public.jmp.com/packages/Principal-Components-on-Correlations/js-p/5c16dba996f7cc0eec8b4bbc