Rodent Little Brother: Secret Lives of Mice

Despite looking very different to mice, our biology is very similar to theirs. Approximately 99% of human genes are predicted to have an equivalent gene in mice, this close genetic similarity means that mice are widely used in genetics research. Mice also have a short gestation time of about three weeks.

Humans have around 20,000 genes and the function of many of these genes is still unknown. What we do know is that many diseases occur when there is an error in a gene. In fact, most diseases have a genetic cause.

Biologists have developed techniques which allow genetic changes to be made in mice. By investigating the consequences of these changes, we can find out more about the underlying processes taking place when a gene goes wrong.

How mice are studied in medical research

Mice are sociable animals and in a laboratory setting they are housed in small groups where they carry out their daily activities such as eating, sleeping, drinking and grooming. In these ‘home-cages’ they can interact with each other and often have their own social hierarchies.

When researchers make genetic changes to mice, they then observe them to understand the implications of the changes. By using different behavioural tests, they try to diagnose the consequences of the genetic alterations. This usually involves taking the mouse out of the cage and putting it into an unfamiliar environment, which can be stressful for the mouse.

The mice are often also observed most during the day, however as nocturnal animals they’re most active at night time, therefore a lot of important data is missed.

The Home Cage Analysis system

Scientists who work with animals are continually striving to refine the way they are studied to advance the welfare of the animals and the quality of the information collected.

As part of the NC3Rs’ CRACK IT initiative to develop a novel system for studying the behaviour of mice in their home-cage, scientists at MRC Harwell have worked with Actual Analytics in Edinburgh to develop the Home Cage Analysis system (HCA).

This system can track the movements of three individual mice without removing the mice from their social group. This method of studying mice is minimally invasive, and only requires insertion of a microchip in a similar way to microchipping a pet cat or dog. The cage is then placed in the Home Cage Analysis System.

Benefits of the project

This system allows mice to be observed undisturbed in their home-cages with other mice. This offers many welfare benefits for how they are looked after including being able to view the mice in the dark (when they are active).

Studying mice in this way also improves the quality of the science. The ability to define early disease characteristics in the home-cage environment will provide us with vital information on the early stages of diseases and potentially how to treat them. Collecting much more data on each individual animal will reduce variation in each experiment and increase the statistical power of testing, so that less animals need to be used.

This project provides an opportunity to gain new knowledge about how mice behave, particularly in the night when they are most active and over much longer periods of time (months and years), which would have been missed if traditional tests were used. It is also possible to observe social interactions in much more detail than was previously possible.

And all the while the mice are going about their daily business, without any human intervention!

We need your help!

"People are great at recognising behaviours in videos, something that is very hard for computers to do. However using lots of examples from people and combining these with techniques called machine learning we can effectively train computers to recognise behaviours from video. Basically they can learn how to do that task automatically by looking at your examples."- Prof. J. Douglas Armstrong,School of Informatics, Institute for Adaptive and Neural Computation. University of Edinburgh.

Collecting data 24/7 is great, but analysing it all is not easy. We aim to use new computer learning algorithms, just as Prof Armstrong has described, to do the work for us. However, for these algorithms to 'learn' what different mouse behaviours are, they first need to be 'taught'. This is done by feeding it examples of such behaviours, these examples are clips of videos that have been labelled by people manually, just like you would be doing in this project.

For the past four years we have had a team of five people doing just that, and have managed to teach an algorithm to recognise when a mouse is climbing.
Now imagine what we could achieve with all of you helping us!

Automated Climbing Annotation:

The project currently creates up to 1 terabyte of data per week. In order have a substantial enough data set to teach and test the computer learning algorithm, we will need to manually analyse this data on a large scale, and without your help it would take us years!

Getting involved is an opportunity to help with a project that has the potential to change the way mice and other animals are studied in the laboratory.

This will ultimately help advance our understanding of diseases and aid the future development of new therapies and treatments.

Find out more

To learn more about the Rodent Little Brother project click here

To read about the science click here

To find out more about how mice are looked after at MRC Harwell click here

To see a 360° tour of the Mary Lyon Centre, our animal facility at MRC Harwell click here

To Read our Mouse Welfare Statement click here