Using My Thai Chili Experiment to Describe Variability of Diversity, Part 2

By Daniel McDonald, PhD

Welcome back to Part 2 of our discussion on diversity! In this blog post, I’m going to talk more about microbiome diversity using a series of samples I collected from myself during a thai chili experiment. We left off the last blog post discussing how both coarse-level taxonomy (i.e., phylum) and within sample diversity can be incredibly variable from day to day in a single person. For example, one of the main observations from the first blog post is that my own microbiome diversity goes from well above the population median to well below the population median in a matter of a few days!


Figure 1 (repeated from Part 1). Alpha diversity histograms under Shannon Entropy and Faith’s PD. The points shown correspond to samples collected prior to and following the chili pepper intervention, where white indicates an early sample and red indicates a later sample. The green line is the median of the background histogram.


You might recall that, anecdotally, I’m sometimes more diverse than the population I’m comparing to. But, what about the diversity of the organisms within individual phyla? To explore this, I’ll focus on Firmicutes, one of the major taxonomic groups found in the human gastrointestinal tract. Despite the fact that humans and other vertebrates evolved alongside this phylum, some fear the Firmicutes, labelling them as “evil.” Please allow me to dispel that myth: demonizing Firmicutes because of a few nasty organisms is analogous to calling the genus Homo evil because of global warming. So, while it is the case that there are some nasty Firmicutes, it is absolutely not the case that Firmicutes are as a whole bad and in fact, there are quite a few beneficial organisms found in that phyla-including those that make delicious things like yogurt!
So, back to my samples. As we saw last week, my taxonomy summary plots show that the abundance of Firmicutes in my samples is similar to the abundance of this phylum in others:


Figure 2 (repeated from Part 1). A phylum level taxonomy summary showing the relative abundances of the top 5 phyla. Others refers to the average relative abundance for everyone not me.

But, I was curious about how diverse my Firmicutes are relative to others? And what about some of the other phyla represented? One way to begin to get at this question is to partition the samples by phylum, and, as before, assess the diversity of each phyla within each sample:


Figure 3. Within phyla diversities. Each row summarizes a different Phylum, sorted by the average relative abundance in the background population. The first column shows Faith’s PD, the second shows Shannon Entropy, and the third depicts the relative abundance of that phyla in the background population; the x-axes are shared within a given column. The green line denotes the median value for each distribution. The colored points show each one of my samples, where white corresponds to the earliest collection time point and dark red to the last collection time point.


What I found interesting is that while I tend to have a “normal” alpha diversity under both Shannon and PD overall (refer to Figure 1), I tend to have a higher Shannon diversity within the Firmicutes but a lower Shannon diversity within the Bacteroidetes. Weird! Well, that got me curious about whether I was dominated by individual organisms within each phyla relative to others. The way I chose to explore this is by using a twist on a classical ecology figure called a rank-abundance plot. In the classic figure, the “rank” of each organism is on the x-axis, and rank is based on the relative abundance of that organism, which is indicated on the y-axis. The twist here is that instead of plotting species on the x-axis, I’m plotting samples, and only illustrating the relative abundance of the most abundant organism in that sample within the given phyla. What’s cool here is that even though the relative abundance of Bacteroidetes in my samples is reasonably high, the diversity (Shannon and PD) were somewhat low, and the rank-abundance curve appears to support this as I tend to have single organisms that have a high relative abundance within the Bacteroidetes (Figure 4 below). That’s pretty cool.



Figure 4. Rank-abundance plot of the five most abundant phyla. The relative abundances are normalized within each phylum (i.e., a relative abundance of 1 means a single organism dominates that phyla). The black dots correspond to my samples.

So, what’s the take home here? As it turns out, there are a lot of ways to both look at and interpret diversity. There isn’t a single “right” way to do it, and this blog post is barely scratching the surface. Ultimately, having a single sample gives some idea of composition of the microbiome, but as you can see, we tend to vary quite a bit day to day even without doing weird experiments on ourselves. It could be the case that my gut was just variable because I was a graduate student at the time, so perhaps the pizza, beer, and limited sleep played a role. Nevertheless, this type of variability is not atypical, which reinforces why scientists love large sample sizes.

In the third and final blog post in this series, I’ll discuss beta-diversity, which is one of the ways we can assess the similarity between samples, and is one of the more powerful measures of diversity.


Daniel McDonald is the former project manager for the American Gut, and is now a bioinformatics scientist at the Institute for Systems Biology focusing on the Wellness 100k Project.

NOTE: The author of this post, Daniel McDonald, is intentionally identifying which samples are his. All participant samples are always de-identified, although participants have the right to identify their sample(s). This decision is solely at the discretion of and under control of the participant.