Skip to main content

Using animal-alternative methods to model organisms

We are using computer-simulated methods (in-silico) to model organisms and better understand neural networks.

nematode under microscope
Developing computer simulations of nematode movement can help us understand neural networks better without using animals.

Biological neural networks are intricate, efficient and effective. We can learn a lot from their traits and use this to develop more complex artificial neural networks. But how do we achieve this without using animals in our research?

At Bath, we’re committed to the replacement, refinement and reduction of animals in research (3Rs). One of the ways we’re doing this is by using computer-simulated methods (in-silico) to model organisms. By developing alternatives to in-vivo and in-vitro research, we can reduce, or even avoid, using animals or animal tissue in experiments.

Modelling the movement of nematodes

We have developed an in-silico model of the locomotion system of the nematode, Caenorhabditis elegans (C. elegans). Using this model, we can better understand the effect of gene mutations on its behaviour. And we can learn more about how segmented neural networks operate.

C. elegans is a transparent free-living nematode that has a total of 302 neurons in its nervous system. Its locomotion system has 86 neurons and it moves backwards and forwards in a serpentine fashion. It has 10 segments along its length, each of which contains a ventral and dorsal nerve cluster as well as several muscles. We have managed to successfully simulate this in our in-silico model.

Read the full article

A simulation success

We implemented the locomotion model using a Field Programmable Gate Array (FPGA). This device enables us to represent the model in-silico and simulate the locomotion system in near real-time. We developed a 2D mechanical avatar of C. elegans to capture any movement when artificial stimuli is applied to the model.

First, we evaluated the model for forwards and backwards locomotion and we were able to compare microscopy of a C. elegans in-vitro with the avatar in-silico. Next, we evaluated how the in-silico avatar behaved when applying a UNC25 gene knockout. This knockout causes the dorsal and ventral neurons to latch into a constant firing mode and the animal appeared to 'seize up'. The in-silico avatar reacted in an identical fashion and seized up. This proved that the avatar not only represents normal behaviour but that we can also use it to predict the effect of specific gene mutations.

‘The results are exciting and promising. It shows we can use in-silico models to simulate an organism. And it brings us another step closer to replacing in-vitro and in-vivo experiments.’
— Dr Ben Metcalfe Department of Electronic & Electrical Engineering

Animal research at Bath

Find out more