Common Lab Mouse Species and Strains Used in Research

Laboratory mice are overwhelmingly one species, Mus musculus, but the strains bred from that species differ so dramatically in genetics, immune function, metabolism, and behavior that choosing the wrong one can derail an entire study. Hundreds of distinct strains exist, each shaped by decades of selective inbreeding or deliberate genetic engineering. Understanding what sets the major strains apart, and why those differences matter, is central to interpreting virtually any mouse-based biomedical finding.

Where Lab Mice Come From

Modern laboratory mice are not descended from a single wild population. Genetic analysis shows that several subspecies within the Mus musculus complex contributed to the genome of standard inbred strains, making their origins polyphyletic rather than neatly traceable to one ancestor.¹ Most of the classical strains trace back to “fancy mice” kept by hobbyists in the early 1900s, which were already hybrids of M. m. domesticus, M. m. musculus, and other subspecies. Researchers then brother-sister mated these animals for twenty or more consecutive generations to create inbred lines, each one genetically near-identical within its strain but distinct from every other strain.

That process produced a powerful experimental tool but also a narrow genetic window. Classical inbred strains capture only a fraction of the variation found in wild mouse populations. A recent effort to address this created eleven new wild-derived inbred strains from wild-caught M. m. domesticus. Each of those strains carries between about 4.7 and 6.5 million single-nucleotide differences compared with the standard reference genome, and over 40% of those variants are absent from classical lab strains entirely.² Domestication has also reshaped behavior: lab strains are significantly less active and slower to explore unfamiliar environments than wild mice.³

Inbred Versus Outbred Stocks

The first decision most researchers face is whether to use an inbred or an outbred stock. Inbred strains are genetically uniform: every mouse within a strain is virtually a clone of every other, which reduces experimental noise and makes results easier to replicate. The trade-off is that any quirk baked into that genome will show up in every animal. Choosing the wrong background strain for a genetically modified line can change the phenotype dramatically, because strain-specific traits interact with the introduced mutation.⁴

Outbred stocks, by contrast, maintain genetic heterogeneity through managed breeding schemes. They are widely used in toxicology, pharmacology, and general screening because their variability better mimics a genetically diverse human population. But that diversity is a double-edged sword. Outbred stocks have often been generated in haphazard ways, and in many studies they are used inappropriately, wasting animals and resources on experiments that would have been better served by an inbred strain or a more carefully characterized population.⁵ Common outbred stocks include CD-1 (also called ICR) and Swiss Webster, both widely available from commercial vendors.

The Major Inbred Strains

A handful of inbred strains dominate biomedical research, each with a distinct metabolic, immunological, and behavioral profile. The most important ones show up across so many fields that their quirks have become common knowledge in animal research facilities worldwide.

C57BL/6

C57BL/6 is far and away the most heavily used inbred strain in the world. It is the background onto which most genetically engineered lines are bred, and it serves as the reference genome for the mouse. These mice are relatively resistant to developing tumors on their own, are susceptible to diet-induced obesity, and tend toward a Th1-biased immune response. In cancer research, C57BL/6 mice show stronger natural killer cell activity and are roughly three times more resistant to melanoma liver metastasis than BALB/c mice.⁶

C57BL/6 mice have a notable metabolic wrinkle, though. The C57BL/6J substrain, maintained at The Jackson Laboratory, carries a mutation in the gene for nicotinamide nucleotide transhydrogenase (Nnt), a mitochondrial enzyme involved in insulin secretion. This results in impaired glucose tolerance and reduced insulin secretion, even though insulin sensitivity remains normal.⁷ Researchers studying diabetes or metabolic disease on this background need to account for that built-in defect.

BALB/c

BALB/c is the second workhorse strain in immunology and cancer research. These mice lean toward a Th2-type immune response, meaning their immune system favors antibody production over the cell-killing responses that C57BL/6 mice excel at. BALB/c mice are commonly used in studies of infectious disease, autoimmunity, and vaccine development. In cancer immunotherapy research, BALB/c mice have been used as the standard host for fibrosarcoma models using WEHI-164 tumor cells, where vaccination strategies can be tested for their ability to stimulate cytotoxic T-cell responses.⁸ Their coat is albino white, which makes them easy to distinguish visually from the black-coated C57BL/6.

DBA/2

DBA/2 holds the distinction of being the oldest inbred mouse strain still in use, established in the early 1900s. It is perhaps best known in hearing research because it develops early-onset, age-dependent hearing loss, particularly at high frequencies, making it a valuable model for studying presbycusis.⁹ DBA/2 mice are also more susceptible to certain cardiovascular and renal conditions than C57BL/6, which makes them useful in studies of those diseases but a poor choice as a general-purpose control.

Why Substrains Matter More Than You Think

Even within a single named strain, substrains can diverge in ways that change experimental outcomes. The most consequential split in mouse genetics is between C57BL/6J (from The Jackson Laboratory) and C57BL/6N (originating from the National Institutes of Health). These two lines separated in 1951 and have since accumulated enough genetic drift that they show significant differences in physiology, biochemistry, and behavior across multiple testing centers.¹⁰

The Nnt mutation mentioned earlier is one dividing line: C57BL/6J carries it, and C57BL/6N does not. But metabolic differences between the two cannot be explained by that one gene alone.¹¹ Researchers who treat “C57BL/6” as a single entity without specifying J or N risk introducing hidden variability into their data. Publications that do not report the substrain make their results harder to replicate and interpret.

Immunodeficient Mice and Xenograft Models

Studying human cancer cells in a living animal requires suppressing the mouse immune system enough that it will not reject foreign tissue. Several immunodeficient strains have been developed for exactly this purpose. The classic nude mouse (which lacks a thymus and therefore has no functional T cells) was among the first, but more severely immunocompromised strains have since taken over for most xenograft work.

NOD-SCID mice combine a defect in B and T cell development with reduced natural killer cell activity, improving the take rate for transplanted human tumors. The most widely used immunodeficient strain today is the NSG (NOD-SCID-gamma) mouse, which adds a knockout of the interleukin-2 receptor gamma chain, eliminating functional natural killer cells entirely. These mice accept human cancer cells, tissues, and patient-derived organoids at high rates, making them the standard host for patient-derived xenograft (PDX) models.¹² Xenograft models using these strains can be ectopic (tumor placed under the skin), orthotopic (placed in the organ of origin), or metastatic (designed to test spread), depending on the research question.

Humanized Mice

Taking immunodeficient mice one step further, researchers can engraft them with human hematopoietic stem cells so the animals develop a functional human immune system. These “humanized” mouse models have become powerful tools for studying human immunology, infectious diseases caused by human-specific pathogens, and for evaluating vaccines and immunotherapies that target human immune cells.¹³ NSG mice are the most common platform for humanization because their deep immunodeficiency allows high levels of human cell engraftment. The approach is not perfect: the reconstituted immune system does not fully recapitulate human immunity, and graft-versus-host disease can develop over time. But for diseases where animal-specific immune responses make standard mouse models misleading, humanized mice are often the best available option.

Disease-Specific Models

Many of the most widely used mouse strains were chosen or engineered because they naturally develop, or can be made to develop, diseases that resemble human conditions. Two fields illustrate this particularly well: metabolic disease and Alzheimer’s disease.

Obesity and Diabetes

The ob/ob mouse and the db/db mouse are both massively obese due to disrupted leptin signaling, but they model different metabolic endpoints. The ob/ob mouse lacks functional leptin itself, while the db/db mouse lacks the leptin receptor. Both become severely obese, but their glucose metabolism diverges sharply: db/db mice develop pronounced type 2 diabetes with elevated glucagon and islet-cell dedifferentiation, while ob/ob mice maintain somewhat better glucose control.¹⁴ Their fat distribution also differs: db/db mice accumulate more subcutaneous fat, while ob/ob mice store more visceral fat and develop worse liver steatosis. The genetic background determines diabetes susceptibility even when the same signaling pathway is disrupted.¹⁵

Importantly, db/db mice capture features of human type 2 diabetes that diet-induced obesity in normal C57BL/6J mice does not. Even when body weight is matched, db/db mice show distinct alterations in metabolic flexibility and islet-cell biology that more closely mirror the human disease.¹⁶

Alzheimer’s Disease

Alzheimer’s research relies heavily on transgenic mouse models that develop amyloid-beta plaques in the brain. The 5xFAD mouse, which carries five familial Alzheimer’s mutations, is one of the most widely used. These mice show progressive amyloid accumulation, synaptic mitochondrial dysfunction, and spatial learning and memory impairments that worsen with age.¹⁷ Newer knock-in models like the APP NL-G-F mouse avoid the problem of artificially overexpressing amyloid precursor protein, which can introduce artifacts unrelated to the actual disease.¹⁸ No single model captures the full complexity of human Alzheimer’s, and each model’s strengths and blind spots shape the kinds of questions it can answer.

Genetic Engineering and Background Effects

CRISPR/Cas9 has transformed how quickly researchers can create new mouse models. The technology works by injecting a guide RNA and the Cas9 enzyme into a fertilized mouse egg, directing a cut at a specific place in the genome. The cell’s repair machinery then either introduces a disabling mutation (a knockout) or, if donor DNA is provided, patches in a desired sequence (a knock-in).¹⁹ Multiple genes can be knocked out simultaneously by injecting several guide RNAs that each target a different gene, and this approach has been shown to work in the first generation of animals without requiring further breeding.²⁰

But the genetic background onto which a mutation is placed can profoundly alter the resulting phenotype. The same gene knockout can produce completely different outcomes, varying penetrance, or different severity depending on whether the animal is on a C57BL/6, BALB/c, or 129 background.²¹ A heterozygous Smad4 knockout, for example, causes overweight in most Collaborative Cross mouse lines, but a few lines actually gain less weight with the mutation present, illustrating how genetic context modifies even straightforward gene effects.²²

The Collaborative Cross and Diversity Outbred Populations

Recognizing that classical inbred strains represent a narrow slice of mouse genetic diversity, researchers developed two complementary resources. The Collaborative Cross (CC) is a panel of recombinant inbred strains derived from eight genetically diverse founder strains, including three wild-derived lines. Using the 63 CC strains, or a population of Diversity Outbred (DO) mice bred from those same eight founders, researchers can map genetic contributions to complex traits with far greater power than would be required in human genome-wide association studies.²³ The DO population is particularly useful for fine-mapping because every animal is genetically unique, mimicking the heterogeneity of a human population while remaining a controlled laboratory resource. These populations have proven especially valuable in immunology and infectious disease research, where host response involves many interacting genes.

Hidden Variables That Change Results

Even when researchers choose the right strain, factors beyond genetics can quietly reshape their data. Two hidden variables have received increasing attention in recent years: the gut microbiome and housing temperature.

Vendor and Facility Effects on the Gut Microbiome

Mice ordered from different commercial suppliers arrive with measurably different gut bacteria, even when they are the same strain. The differences between a C57BL/6 mouse from one vendor and a C57BL/6 mouse from another can actually outweigh the differences between two entirely different strains purchased from the same vendor.²⁴ These are not trivial differences. In one study, mice injected with fecal matter from different vendors developed dramatically different severity of sepsis: stool from some vendors induced lethal sepsis, while stool from another vendor produced no symptoms at all.²⁵ Substantial variation in bacterial composition has been documented even among mice from the same vendor shipped to different facilities, and that variation can shift further after the animals spend time in a new vivarium.²⁶

Cold Stress From Standard Housing

Most mouse facilities keep room temperatures between 20 and 26 °C, following standard guidelines. But mice are comfortable at 29 to 34 °C. That gap means lab mice are chronically cold-stressed, spending extra energy to maintain body temperature. This raises their baseline metabolic rate, shifts their metabolism toward burning more glucose, and activates brown fat tissue that would otherwise be quiescent.²⁷ The consequences extend beyond metabolism. Cold-stressed mice mount suppressed anti-tumor immune responses because their bodies are diverting resources toward thermoregulation. Tumor growth rates and the effectiveness of immunotherapy look different in mice housed at thermoneutral temperatures versus standard vivarium temperatures.²⁸ This means that decades of cancer immunology data were generated in animals whose immune systems were operating under chronic physiological stress that humans do not typically experience.

Sex as a Biological Variable

For years, many preclinical studies used only male mice to avoid the perceived complication of hormonal cycles in females. Funding agencies, including the NIH, now require that sex be considered as a biological variable in study design, because disease presentation and treatment responses can differ between sexes in ways that matter for human medicine.²⁹ Specialized mouse models have been developed to tease apart the contributions of gonadal hormones from those of sex chromosomes. The Four Core Genotypes model and the XY* model allow researchers to separate chromosomal sex from hormonal sex, revealing effects that would be invisible in standard male-versus-female comparisons.³⁰

The issue plays out concretely in Alzheimer’s research. The 5xFAD model uses the Thy1 promoter to drive its transgenes, and there is evidence that Thy1 promoter regulation is sex-dependent, which could skew phenotypic outcomes in ways that masquerade as genuine sex differences in disease biology.³¹ Investigators using any transgenic line driven by Thy1 need to be cautious about attributing sex differences in their results to the disease rather than to the construct.

Beyond Mus Musculus

Not every rodent model in the lab belongs to the house mouse species. Spiny mice (Acomys spp.) are a striking example of a non-standard rodent earning a growing role in research. Unlike typical mammals, spiny mice can regenerate skin, ear tissue, and even cardiac muscle after injury, deploying sustained pro-regenerative pathways while keeping fibrotic scarring transient.³² Understanding how these animals avoid the default mammalian scar-forming response could eventually inform human wound-healing therapies. Other rodent species used in specialized research include the Mongolian gerbil (epilepsy and hearing research), the deer mouse Peromyscus (aging and altitude adaptation), and various vole species (social bonding and monogamy). Each fills a niche that Mus musculus strains cannot, and their increasing availability reflects a broader recognition that one species, however well-characterized, cannot model the full range of mammalian biology.