BuiltWithNOF
Strategy in use of inbred strains

An inbred strain represents only a single genotype.  Generalisations based on such material must be treated with caution,  although in-depth studies of a single strain have resulted in a major scientific breakthrough on several occasions. For example, work on the  BALB/c mouse strain led to the development of monoclonal antibodies, and the  development of embryonic stem cells and the technology for "knocking out" individual genes depended on strain 129.

Inbred strains may be used in several ways, which are  discussed below.

Many scientists use a single inbred strain or F1 hybrid in their research because it has a repeatable, standardised, uniform genotype supported by a substantial body of previous research. Thus, between 2001 and 2005 over 20,000 scientific papers used BALB/c mice and a further  10,000 used C57BL/6 (see Table 2 of the Mouse strains page). In many ways, the more that is known about a strain the more valuable it becomes.

The phenotypic uniformity of an isogenic strain means that sample  size can be reduced in comparison with the use of outbred stocks. The main disadvantage of using a single strain in this way is that it is never clear whether the results will generalise to other strains.  However, in most cases it is sufficient to note the strain of mouse or  rat in the title or abstract.

Use of a single isogenic strain as standard experimental material

 The expression of most mutations (including transgenes and knockouts) is  dependent on the genetic background strain. An outbred background such  as "Swiss" or CD-1 is not ideal because it will result in a more variable phenotype which will be subject to genetic change as a result of inbreeding and directional selection.

Thus, it is standard practice to backcross mutations to an inbred  background. This carries some risk as the phenotype may become more  severe on some backgrounds to the extent that mutant animals may survive  well with some backgrounds, but not on others. So it may be necessary to backcross to more than one inbred strain. However, if the expression of a mutation does depend to a marked extent on the genetic background,  this implies that there are modifiers of expression which may themselves  be or interest.

When a mutation has been backcrossed to an inbred background for at  least ten generations it is said to be congenic with the  background strain. A pair of congenic strains will differ at the  differential locus and an associated piece of chromosome but should  otherwise be genetically identical. Thus the background strain becomes  the control for the mutation in any comparative studies.

Use of an  isogenic strain as the genetic background for a mutation or transgene

Some investigations require the transfer of cell,  tissues or organs between animals. Cells or tissues transplanted between isogenic individuals will not be immunologically rejected. F1 hybrids  between two inbred strains will also accept grafts from both parental  strains.  In most cases it will also be possible to transfer tissues between congenic strains provided they do not differ at loci associated  with graft rejection.

Where graft rejection itself is being studied it is almost essential to  have a known, controlled rate of rejection. This can best be achieved by  choosing a pair of inbred strains known to differ either at the major histocompatibility locus, or at minor loci, depending on the desired speed of graft rejection.

Use of one or more strains because inbred strains are  essential to the study

 Several strains are used almost exclusively because they get some disease or  have some useful characteristic not found in other strains. NOD mice get  type I diabetes, NZB mice get autoimmune anaemia, AKR mice get leukaemia and SHR rats are hypertensive and WAG rats get absence epilepsy. These strains are used largely to study these conditions. Strains with useful properties include 129 because it is the strain which has been used to generate  knockout mice, and strain FVB because one-cell embryos have a large male  pro-nucleus into which it is relatively easy to inject foreign DNA.

A particular strain may also be chosen for further study because it is susceptible or resistant to a pathogen or a toxic chemical. Strain A mice are susceptible to the induction of lung tumours by carcinogens and to cleft  palate by teratogens. Strain C57BL mice have a preference for alcohol.

A strain may also have some characteristics which make it unsuitable or less suitable for a particular investigation. Clearly, toxicologists do not want  to use strains such as AKR which have a short lifespan due to a particular  disease. The males of strains BALB/c and SJL can be highly aggressive so  that fighting may occur between males housed together. These things need to  be taken into account when choosing a strain for a particular application.

Use of a strain because it has a disease or some other interesting characteristic

 In some cases the use of a single inbred strain or outbred stock can give misleading results. This is particularly the case in toxicity testing where  a strain or stock may be resistant to a particular chemical. For example, outbred Sprague-Dawley rats appear to be resistant to levels of  diethylstilbestrol (a known human carcinogen), which give rise to many tumours in other strains. Example 3 in the multi-strain pages shows strain differences in response to a chemical causing prostate tumours in rats.

The use of several strains using a factorial experimental design, which does not increase the total number of animals, can show the extent to which it may be possible to generalise the results.

In most cases four or five strains will be sufficient to give a good general idea of the extent of any genetic variation in response.

Use of two or  more strains to increase generality

 Strain differences in spontaneous disease, behaviour, anatomy, physiology and response to pathogens and chemicals are usually due to more than one genetic locus plus environmental factors. There is now a substantial amount of work going on to try to identify the genes involved. This usually means selecting two strains which differ, crossing them, and mapping the genes which contribute to those differences. The most important of these may then be identified, although this last step is still quite difficult. Once these  QTLs have been identified the homologous genes in humans can be found and  any contribution to disease investigated.

Use of two strains to identify QTLs (quantitative trait loci)

 Unusual strains are often of particular interest. They may have a rare  disease or physiological or neurological condition which would reward  further study, or they may just be the extreme strain in a continuum. The mouse and rat phenome studies are characterising many strains on a broad  range of phenotypes, and can be expected to show up any strains which are unusual for a particular character, together with any identified gene loci that contribute to the trait.   However, there are an infinite number of characters of potential interest, and much work remains to be  done. Anyone doing a comparative study of this sort should ensure that the data are acceptable to go into a relevant database.

Strain surveys to identify interesting strains

In many cases it is desirable to know whether two characteristics are  related. Is the level of a hormone related to a specific behaviour, or is the rate at which a strain metabolises a drug related to sensitivity to that drug?. If several inbred strains are used  it is possible to discover whether or not two characters are related. Note that a correlation coefficient, which might be used for comparing two or more characters across strains has N-2 degrees of freedom. Thus, this approach requires a  minimum of three strains, and only has any real power with five or more strains. It is not appropriate to take two strains, measure character A and B in each, and then claim that because  characters A and B are both high in one strain and low in the other, this implies some association between the characters. This approach is  particularly tempting where there is a strain with a character of particular  interest, and a "control" strain such as the hypertensive rat strain SHR and  its "control" WKY. Research workers have been tempted to assume that any difference between these two strains (such as activity) must be associated  with hypertension. The use of three or more strains should help to reduce false claims of this sort.

Use of several strains to show any relationship between characters

In many cases it is desirable to know whether two characteristics are  related. Is the level of a hormone related to a specific behaviour, or is the rate at which a strain metabolises a drug related to sensitivity to that drug?. If several inbred strains are used  it is possible to discover whether or not two characters are related. Note that a correlation coefficient, which might be used for comparing two or more characters across strains has N-2 degrees of freedom. Thus, this approach requires a  minimum of three strains, and only has any real power with five or more strains. It is not appropriate to take two strains, measure character A and B in each, and then claim that because  characters A and B are both high in one strain and low in the other, this implies some association between the characters. This approach is  particularly tempting where there is a strain with a character of particular  interest, and a "control" strain such as the hypertensive rat strain SHR and  its "control" WKY. Research workers have been tempted to assume that any difference between these two strains (such as activity) must be associated  with hypertension. The use of three or more strains should help to reduce false claims of this sort.