Finally, after years of blood sweat and tears our comparative connectomics paper has been published. This represents a huge victory for myself and all of the people who have helped me along the way. It is rather humbling to have this work get the kind of exposure it gets by being published in Cell. It is the culmination of years of work, complete with with heart-breaking failures, long grinding hours on the microscope and in front of the computer, and the occasional adrenaline-inducing discovery. In the paper, we compare a wiring diagram(or connectome) of the pharyngeal nervous system of the nematode Pristionchus pacificus to that of the well-known model organism C. elegans. The data are obtained by slicing plastic-embedded animals into ultrathin sections (about 50nm each) and imaging them under the electron microscope. In pursuing this work, we generated 3 data sets, each over 3000 sections long. Imaging resulted in nearly 200,000 individual exposures of the camera, and all of the challenges that result from organizing and annotating that many images.
So why compare the two species? It turns out that there are some cool things about nematode nervous systems. Generally, cell identity in these worms is extremely highly conserved. I knew this when I started the project, but I was still surprised to find that in the pharynx there is exactly the same number of neurons (despite some ambiguity in published literature) and that for every single neuron I could find a clear counterpart in C. elegans! This, despite the fact that they are not at all closely related! They may have even diverged at a similar time that the human lineage diverged from sea urchins, so they are really distant cousins with identical building blocks for their nervous systems. In terms of the graph theory, this conservation in cell identity gives us the power to compare networks where all of the nodes are the same and it is only the edges that differ. Another reason to compare them is that their behavior is markedly different. While C. elegans is a bacterial feeder, P. pacificus can switch its behavior to take advantage of multiple food sources. Most interestingly, it can be predatory on other nematodes! So we essentially have the same nervous systems but different behavior. Would we be able to see these behavioral differences reflected in the synaptic connectivity?
Turns out, we can. Despite conserved cell identity, there is a massive difference in synaptic wiring between P. pacificus and C. elegans. In fact, fewer than half of the synapses found in one species could be found in the other. Furthermore, many of these differences seem to be associated with behavior that exists in one species but not the other. Cool, eh? This was not necessarily the expected result. In fact, most evolutionarily changes in behavior that have been documented are the result of modulatory or physiological properties of neurons, rather than being explained by new or removed synapses.
My background is in ultrastructural anatomy in nematodes, so I was unusually well prepared for generating these datasets. But at some point we had a bunch of data and had to analyze it somehow. For this, I turned to the unfamiliar territory of Graph theory, partly inspired by some pioneering work by Dmitri Chklovskii and his group. I found some graph theory to be useful… specifically centrality metrics, which ask the question “What are the most important nodes in the network?”. But as biologists, we very often have to answer the inverse question, or “Which parts of the network are important for my node?”. For this, I failed to find acceptable methods. I came up with a modified version of centrality, calling it focused centrality, that helps answer that question by employing an explicit model of information flow and by only considering portions of that information for whatever specific question I had. I worked very closely with Christian Rödelsperger, our resident bioinformatitian, to develop the idea and produce some software to implement it. In the end, I think we have found a focused centrality approach to be a valuable objective way to describe how the network appears in importance to a particular node. We’ve got plans to ask some additional sorts of questions with it in future work as well. In our paper, we use it to highlight differences in connectivity for nodes that are likely involved in predatory feeding in P. pacificus.
Enough words…. The paper can be found HERE. If anyone ends up here and has any questions or comments, feel free to get in touch!