How model organisms are chosen


Mice, fruit flies, worms, and monkeys. These are just some of the many animal species that are commonly used by scientists to learn about the nervous system, often with the goal of uncovering something about the human nervous system. If that’s the case, then the species being used is said to be a model organism, since it is intended to model some aspect of human biology. But in the hundreds of years since the scientific revolution in the 1500s, how did we settle on a relatively small handful of species to study? There are 8.7 million different species out there – why were these species chosen and not others? How much can they really tell us about human biology? 

Fig. 1 Some of the most popular species to use as model organisms in research labs around the world. From left to right: African clawed frog, microscopic worm, zebrafish, axolotl, naked mole rat, mouse, fruit fly, and cuttlefish. Illustration by Charlotte Hu, Cold Spring Harbor Laboratory

Why use model organisms?

For starters, the necessity of model organisms stems from the fact that understanding human biology is largely limited to studies of post-mortem samples, observation (e.g. MRI, ultrasound imaging, etc.), and in vitro studies of human cells (i.e. cells on a petri dish), the latter of which are still an imperfect model of the complex human body. However, there are several questions that can only be answered by invasive manipulation of the organism being studied, or very detailed autopsies at specific time points. Let’s say for example that you want to know if a particular gene is involved in the development of multiple sclerosis (MS), a disease caused by an autoimmune attack of the nervous system. To definitively prove this in humans, you would have to delete this gene in people, and either observe to see if they develop MS, or try to give them MS and see if they are unable to develop it. However, deleting this gene might permanently cause terrible unforeseen effects to their overall health. The moral problem of this is compounded by the fact that scientists always need to repeat their experiments many times to be sure of their results, meaning many people would have to suffer in this scenario. 

As such, experimenting on humans, especially experiments requiring invasive procedures, is rife with ethical concerns, and is only done once a treatment has been shown to be safe in animals. Even then, these procedures are limited to individuals who would benefit from said experimental treatments (i.e. those who are sick) in order to avoid giving an unnecessary medical intervention in healthy people. 

However, these questions are still important because answering them could improve the lives of many people. So in order to answer them, but cause no harm to humans, scientists turn to the use of model organisms. This isn’t to say that scientists don’t care about the well-being of the organism they choose. By law scientists must ensure that the organisms they use to study are well kept and suffer the least amount possible throughout their lives.

With so many species to choose from–how do you pick?

Normally, researchers select the best species to use for their experiments based on certain criteria. In no particular order, these criteria include (but are not limited to):

  1. Relevance to question or disease being studied
  2. The ability to produce a lot of offspring in each generation (fecundity)
  3. Availability of tools, knowledge, and resources to conduct high-level experiments.(e.g. sequenced genome, transgenic strains, veterinarians trained to care for these organisms)
  4. Cost of maintaining a colony of organisms and the ease to which they can be cared for 
  5. Time it takes for the organism to mature or be relevant to the question at hand

The first criteria is a given. If you want to study complex learning and memory, it doesn’t make sense for you to use yeast cells, which are single-celled organisms, as a model. An animal  like a mouse or a monkey which can be trained to perform certain recall tasks would be more suited to the question. Points 2- 5 may seem less obvious, but are very important to a researcher who is working on a limited time and financial budget. 
It would be fantastic to learn why elephant brains look so normal into old age [1], but feeding, housing, and caring for many families of this endangered species would be costly, take a long time, and be overall impractical for a lab setting. On the other hand, C. elegans worms, which require nothing more than some food on a petri dish, quickly multiply, and whose biology we know a lot more about, can be a greatly pragmatic model to understand very fundamental questions about neuroscience. You can read more about why C. elegans worms are a great tool to use in neuroscience research here.

Evolutionary considerations

If your research goal is to understand a human neurological disease, then your best bet to get the most accurate model of a human disease is to get something as close to (evolutionarily speaking) humans. In the animal kingdom, that would be other primates such as bonobos or chimpanzees, which share about 99% of our DNA [2]. Even though some labs around the world, such as the Primate Research Institute in Kyoto, Japan, do study the cognitive skills of these animals, they are still challenging and expensive to care for given the necessary size of their enclosures and their endangered status. As a result, some labs have opted to use monkeys like the Rhesus macaque or the marmoset instead, which are on average smaller, reproduce more quickly, and are not on the endangered species list making them a more practical (and ethical) option. The drawback is that they are further back in the evolutionary tree (macaques share about 93% of their DNA with us) and are still decently expensive to care for [3]. 

Traveling further back in the evolutionary tree, we find mice and rats. These rodents reproduce very quickly, produce large litters, don’t require as much space, and have a tremendously large amount of strains used to study different genes and diseases. The caveat here is that mice, while still being mammals like humans (and sharing about 70% of the protein coding sequences of our DNA), have a substantially different physiology than us [4]. In the fight to choose the best model organism for human biology, the mouse seemingly came atop for being the most convenient and practical mammal to house in a laboratory setting.

Fig. 2. Evolutionary tree of common model organisms. The organisms closest to humans are most similar to us, with the numbers indicating the latest point of divergence in millions of years ago. John Hatch


There is no perfect model organism. Each will have some limitations, but we still use them in science because they are what makes the most sense. And this is by no means a settled debate. There has been a recent outcry to try to move away from the laboratory mouse as the “standard” mammalian model because it is so difficult to study certain diseases in them [citation]. Mice infamously have a strong resistance to developing Parkinson’s and Alzheimer’s disease, where most research labs introduce 4 different mutated human genes only for the mice to kind of get Alzheimer’s. And even then, how do we know what Alzheimer’s looks like for a mouse when they can’t communicate their experience with typical Alzheimer’s traits, such as memory problems? Social and behavioral disorders are even more difficult for us to interpret in a different organism (have you thought about what a mouse with autism might look like?). There are clues that point to us when something is “different” or “wrong”  in an animal when modeling a human disease, but it is difficult to be sure that those differences are genuinely the disease you’re trying to model. As a consequence, many treatments that “work” in mice, fail to work in humans. Referencing this exact phenomena, Thomas Insel, a former director of the The National Institute of Mental Health said, “if you’re going to get Alzheimer’s, first become a mouse.”


A lot of great progress has been made using the model organisms at our disposal, and they are still valuable tools to answer many important questions. Mice, worms, monkeys, and humans are all very different creatures, but some aspects of all our biologies are shared. As such, we can learn a tremendous amount about ourselves by studying them. However, they are, like all things, imperfect, and it is important to be mindful of their limitations as we read up on the latest groundbreaking discoveries or conduct research ourselves. As the great statistician George Box said, “All models are wrong, but some are useful.”


  1. Cole G, Neal JW. The brain in aged elephants. J Neuropathol Exp Neurol. 1990 Mar;49(2):190-2. doi: 10.1097/00005072-199003000-00012. PMID: 2307983.
  2. Waterson, R. H., Lander, E. S., Wilson, R. K. & Consortium, T. C. S. and A. Initial sequence of the chimpanzee genome and comparison with the human genome. Nature 437, 69–87 (2005).
  3. Spencer, G., 2007. Analysis of Rhesus Monkey Genome Uncovers Genetic Differences With Humans, Chimps. [online] NIH News Releases. Available at: <; [Accessed 14 September 2022].
  4. Benowitz, S., 2014. New comprehensive view of the mouse genome finds many similarities and striking differences with human genome. [online] NIH News Releases. Available at: <,1.5%20percent%20of%20these%20genomes.&gt; [Accessed 14 September 2022]