Connectome explorer in the browser

Model Organism: Worms WormAtlas
1

Hi,
I thought some members of this forum might be interested in a little tool I’ve recently built that allows for visual exploration of the c. elegans connectome in the browser, based on the dataset by Varshney et al. (as linked on the neuronal wiring page of the Atlas):

https://elegans.herokuapp.com

In its current form it would be useful primarily for researchers interested in identifying particular subcircuits, hypothesised to be underlying a certain behaviour. E.g. if one were interested in salt klinotaxis, one may wonder about the circuit connecting relevant sensory neurons (ASE) to certain motor neurons (say, SMB). This is what the tool is designed to do. One may select two separate sets of neuron groups to identify the circuit connecting them, while filtering the contained connections according to the number of synapses made and limiting the maximum path length (the number of interneuron “jumps”).

I have plans for adding more ways of querying the connectome (as well as adding sensory features and muscle connections), but I’d be curious to know if there are people who might actually want to use such a tool before putting more work into it.

Also, the tool is based on an earlier Mathematica version by Eduardo Izquierdo, who has used it just as described above to investigate salt klinotaxis circuits (published here).

Cheers,
Thomas.

2

Looks interesting, though it doesn’t quite work in my browser (Chrome) - the instructions text hovers in front and I haven’t figured out how to dismiss it, or even to scroll to the bottom.

People interested in your tool may also be interested in Nikhil Bhatla’s presentation of the C. elegans neural network. It lacks the global view and the 3D representation you have, but has some other useful features.

3

Looks interesting! I get the same problem as Hillel using Firefox and Safari browsers

4

Thanks for the feedback. It might not be too obvious but at the bottom of the introductory text panel is a little handle with a chevron symbol. Clicking that moves the panel out of view. I should probably add a more obvious way to get started…

5

I see. I’m using a macbook with a small screen, so the lower part was cut off the bottom of the screen, and for some reason I couldn’t scroll down to the bottom. However when I decreased the zoom in my browser to fit it in I was able to click on the chevron and get started.

Looks like a useful and simple way to look at the connectivity between neurons, when they’re more than one connection apart. I like how you’ve split the VAs VBs etc into their individual parts, and split the other neurons into L and R. Being able to scale synapse weight and path length is also useful.

Muscle connections would be great, and if you have information to hand on which muscle cell is innervated (e.g. mu_bod (DL1), or mu_bod (VR5)) that would be even better, but of course that depends on if you have the information and want to spend the time.

Thanks for sharing! I would find it useful

Ben

6

Thomas,
I just went and visited the site - very nice and very easy to use. Thanks!
Janet

7

Hi, thanks again for the interest.

It seems showing muscle innervations and sensory receptors would be the most useful feature to add, so I’ll do that as soon as I find the time, which won’t be for a couple of weeks.

Cheers,
Thomas

8

As promised the connectome explorer now includes both sensory receptors and muscles. Since this complicates the querying for subcircuits a little, perhaps a short explanation may be useful.

The primary interaction with the explorer is via the subcircuit panel. Here one can choose two sets of neuron groups to view the network connecting them. The input box for each group has an auto-complete feature, so as soon as you type it will suggest corresponding group names. Hitting any arrow key will bring up a list of all groups. Once a group name has been selected the form will allow you to add the next group name by comma separating them. You can add as many group names as you like, though performance of the search query can take a hit if too many strongly-connected neurons are selected.

Between the two group inputs there is another button to select the directedness of the connections returned between the two groups (uni- or bi-directional).

In addition, there is now a field to select general sensory modalities to include in the circuit. The input mechanics are similar to the group names. The difference being that for each selected sensory modality, all sensory neurons with that modality are now implicitly added to the first group of neurons selected above (which can therefore be left empty now too). Also, information about sensory modality and the function a certain sensory neuron is thought to be involved in is now displayed in the Info panel.

Also, muscles can now be visualized. Selecting any of the head/neck/body tickboxes will include muscles from the corresponding section to be included, if(!) they are directly innervated by any of the neurons in the subcircuit that results from the previous search options (so it’s not (yet) possible to only select a sensory modality and a muscle group to find, say, the circuit connecting nociceptors to muscles in the head; one still has to explicitly include desired motor neurons).

The sliders should be self-explanatory. They select the minimum number of synapses between two neurons that is required to include a connection in the subcircuit, as well as the maximum path length between the two neuron groups.

Finally, after retrieving a specific circuit via the Fetch button, an Expand button becomes available. This will add all existing connections between neurons in the identified circuit, irrespective of the weight filters. Note that this results in different circuits than running the query with filters set to 0 from the start.

In any case, I hope you’ll find the additions useful. Feel free to suggest further features you’d like to see.

Cheers,
Thomas.