8 August 2024

Matthew Michalska-Smith

This talk was given at the Ecological Society of America's annual meeting in Long Beach, CA USA on 8 August 2024. Text here will supplement slide material.

You can remove this pane (and get it back) by tapping v on your keyboard.

Some of the slides will have animations which require advancing the slide partway through these explanations. These will be indicated by "next slide" in these notes.

Slides can be advanced by tapping to display only the core slides, or Spc, PgDn, or n to also show supplemental slides present in some presentations.

Finally, this presentation will look best on a desktop/laptop full-screen, but should be more or less intelligible on smaller screens (no guarantees about phone screens). So, whenever you're ready...

next slide

Linda Kinkel

Molly Kuhs

Zoe Hansen

Julia Ahlborn

SICA: sub-inhibitory concentration of antibiotic
Antibiotic: chemical compound used to inhibit other microbes

Davies, Spiegelman, & Yim 2006

Tibúrcio et al. 2022

SICA: sub-inhibitory concentration of antibiotic
Antibiotic: chemical compound used to inhibit other microbes

chemical compound used to inhibit or signal other microbes

Schlatter & Kinkel 2014

20 Isolates collected from 2 nutrient treatments

Niche Width: the number of resources consumed
Total Growth: summed growth across resources
Growth Efficiency: average growth across resources

Repeated for subinhibitory concentrations of 6 antibiotics

So, with that background, we can now jump into the experimental data that was collected for these analyses:

We collected 20 isolates, 10 from each of a high and low nutrient plot in the Cedar Creek Ecosystem Science Reserve in MN. These nutrient treatments continue today and had been in place for nearly two decades before our isolates were collected, so we can consider these as communities of isolates that have co-evolved under wildly different resource environments.

next slide

Each isolate was then grown on a Biolog plate, which is a 96 well plate where each well contains a distinct carbon source as nutrition.

next slide

We measure the isolate's growth on each of these carbon sources to produce an approximation of each isolate's resource niche, which we can then summarize using different metrics, such as

next slide

the number of carbon sources an isolate grows on,

next slide

or the total growth across carbon sources

next slide

or the average growth across utilized carbon sources.

next slide

Finally, we repeat this procedure six more times for each isolate, this time also adding an isolate-specific subinhibitory concentration of one of six different antibiotics to the wells prior to observing their growth.

Kuhs, et al. in prep

And, indeed, we see effects of SICA on our metrics of isolate growth. In general, SICA reduce total growth, especially in the high nutrient treatment, indicated by many of the boxplots falling below the dotted line indicating the mean for that metric in the control treatment, and the preponderance of red dots over blue ones, with the dot color indicating the direction of change from control for each specific isolate.

Note, however, that there are exceptions to the above. Some isolates show remarkably increased growth under SICA conditions, not just for particular carbon sources, but on average across all 95 sources. Also note how variable responses are across SICA and Nutrient treatments -- we are really just scratching the surface for how SICA affect isolate resource utilization.

Kuhs, et al. in prep

We can also look at qualitative changes in isolate growth across resources. Here, we look at each isolate-carbon source combination and classify it's change from the control condition. Isolates can either stop growing on something they can grow on in control ("Suppressed"), start growing on something they did not in control ("Induced"), or have a more qualitative change in growth, growing more ("Increased"), less ("Decreased"), or the same ("Unchanged") as they did in control.

We see that most isolate-carbon source pairs do experience changes in growth, with more reductions than additions. But this is not the case for all antibiotics, Streptomycin, for instance results in more increases than decreases in the low nutrient isolates. Also note that, inductions, while occasional, are very rare.

min(

)

min(

)

Average across all resources for each (ordered) isolate pair

Taking this one step further, we can use our inferred resource niche to gain insight into inferred apparent competition between a pair of isolates.

next slide

By comparing growth rates on each of the 95 nutrients, for each nutrient and each ordered isolate pair, we can,

next slide

calculate a metric that quantifies how much of the blue isolate's growth will be under competition with another isolate, shown here in red. In other words, if the red isolate grows more than the blue isolate on this resource, then the blue isolate experiences 100% overlap. Importantly, however, this metric is not symmetric.

next slide

considering the view from the red isolate's perspective, we see that only a bit more than a half of the red isolate's growth is under competition from the blue isolate.

next slide

we average these values across all 95 nutrients to calculate a single measure of the blue isolate's overlap on the red isolate, and vice versa.

next slide

If the average value is great than 75%, we consider it a "significant" degree of overlap, meaning we think this level of interaction is biologically meaningful. These values serve as the foundation for a network of inferred, apparent competition

next slide

with these two isolates just forming a small component of the larger, more complicated web of microbial interactions.

Hansen, et al. in prep

We can also consider possible changes to global network structure. Here I plot the values for each of three metrics of network structure: the number of edges in the network, modularity, and the clustering coefficient. Modularity can be thought of as the preponderance of nodes in the network to form groups that interact more strongly within the group than with members of other groups and clustering is a measure of how likely your interaction partners are to also interact with one another. When plotted this way, it can be a bit overwhelming what is going on, so to simplify, we can normalize by the control values

next slide

such that the control value is now zero for each metric in each nutrient treatment. We now see some intesting patterns are present: SICA tend to reduce network connectivity regardless of nutrient condition, but for modularity there is an interaction such that SICA increase modularity above the control condition in low nutrient conditions, but decrease it in high nutrient conditions. For clustering, in low nutrient conditions, SICA networks tend to look like the control, but in high nutrient conditions, SICA networks consistently have lower values.

In Summary,

SICA usually, but not always!, reduce growth rates
Most pairwise interactions change, but on average changes balance out
Some network structure measures vary under SICA, but more idiosyncratically

Linda Kinkel

Molly Kuhs

Zoe Hansen

Julia Ahlborn

Davies, Spiegelman, & Yim 2006

Chow, Ghaly, & Gillings 2021

Lorian 1975

Tibúrcio et al. 2022

Schlatter & Kinkel 2014

Kuhs, et al. in prep

Kuhs, et al. in prep

Hansen, et al. in prep

In Summary,

Thank you!

/

Questions?

Slides available on Michalska-Smith.com