If strain differences are found in a multi-strain study this implies that the outcome is genetically determined. Further action depends on individual circumstances. For example, in a screening experiment there may not be any great need to study any differences in detail, although the reasons why strains differ is always of general interest. However, in some cases the reasons for the strain differences could be of great importance in assessing relevance to humans. Two situations are discussed briefly below.
Use of a multi-strain study to find the most suitable strain for future work
Multi-strain studies can be used to find one strain which is “susceptible” or in some way more suitable than other strains for further study. Research can then be concentrated on this single strain. This is a perfectly acceptable strategy but the findings may not generalise to other strains. The strain used should be noted explicitly in the title or abstract of the paper.
Pharmaceutical companies often use animal models for screening potential drugs, and may carry out essentially the same assay many times with slightly different compounds. If they can find a highly sensitive strain of mice or rats, then sample sizes can be reduced or the power of their screens will be enhanced. A multi-strain study may, therefore, be used as part of the process of optimising such an experiment, with the most sensitive strain being chosen for the screen (see Shaw R, Festing MF, Peers I, Furlong L. 2002. Use of factorial designs to optimize animal experiments and reduce animal use. ILAR J 43:223-232., available on-line at ILAR web site).
Use of a multi-strain study to find loci controlling a character of interest
Strain differences may be due to a single genetic locus or, more usually, to multiple loci with environmental influences, i.e. the character has a polygenic mode of inheritance. There are several approaches to identifying the loci involved.
The first step is usually to see if the strain difference is due to a single gene. If several strains have been phenotyped, and they fall into two distinct classes, or one strain is very different from all the others, then this may suggest that the difference is due to a single Mendelian locus. In this case crosses of a “positive” x “negative” strain can be made. The F1 hybrid should then be phenotyped to see whether there is a dominant/recessive mode of inheritance. The F1 hybrids can then either be inter-crossed and the progeny phenotyped to look for the 3:1 or 1:2:1 Mendelian ratios, or they can be backcrossed to the recessive strain (assuming a dominant mode of inheritance) to look for the Mendelian 1:1 ratio. Tissue samples or DNA should be saved from all animals.
If Mendelian ratios are found, then the next step is usually to map the chromosome location of the locus. If two common inbred strains have been used then data on many DNA-based genetic markers such as microsatellites and single nucleotide polymorphisms will be available on various web sites. Basically, the method is to choose 3-4 of these marker loci per mouse chromosome at which the parental strains differ and type all the F2 and/or backcross animals at these loci. All these markers can be typed using PCR-based methods. Any association between a marker locus and phenotype will be indicative of genetic linkage.
Once an approximate chromosome location has been found, larger numbers of animals and markers on that chromosome will be used to fine-map the gene. The final step is to identify the polymorphic gene. There may be some good candidate loci in the chromosomal region, or it may be necessary to clone and sequence the gene. At this stage, expert advice is usually advisable, as there may be several different approaches which could be considered, depending on circumstances.
If several inbred strains have been phenotyped and they do not seem to fall into distinct groups, but have a range of “susceptibility”, then the character probably has a polygenic mode of inheritance. In this case the aim is to identify the Quantitative Trait Loci or QTLs that control the character. The methods used are very similar to those described above. Usually two strains with the highest and lowest “susceptibility” are chosen, and the analysis aims to map the loci as which they differ. As above, the strains are crossed to produce a backcross or F2 generation and all animals are phenotyped and typed at marker loci at which the parental strains differ. Again, any association between a genetic marker and phenotype is suggestive of genetic linkage, but in this case there may be linkage to several chromosome locations.
The main problem with the QTL analysis is going from map location to the actual locus, so expert advice is essential.
A whole range of specialised strains have been developed which may help in the genetic analysis of a character. Sets of Recombinant inbred strains (RIS) can help to determine whether a character is due to a single genetic locus, and if so it’s approximate chromosomal location. Sets or recombinant congenic (RCS) strains have been developed to help in the genetic analysis of complex traits. Similarly, sets of consomic strains in which a whole chromosome has been substituted for one of a different strain can indicate which chromosomes contribute polymorphic genes to which characters. These strains are discussed in more detail under the Isogenic strains page.