that they have thought for a long time that they could control genetics of the colony through careful breeding schemes. However, as shown in this study, caesarian section produced a bottleneck effect on Jcl:Wistar, and Iar:Wistar is genetically different from Crj:Wistar. Thus, subcolonies exist. The only method for discriminating subcolonies from each other is by genetic testing.

For closed colonies, we propose a monitoring method as follows:

A. Testing methods

The method of the ICLAS Monitoring Center is recommended. We should use the same items for monitoring to facilitate genetic evaluation of each colony.

B. Monitoring procedures

• Genetic profiling may be performed once at the beginning of periodic monitoring and should be repeated every several years.
• Monitoring profiling should be carried out periodically (for example, once a year) using a set of markers selected among the genetic profiling markers shown in Table 2.
• Testing numbers should be done, randomly selecting from a production colony with a requirement of 50 to 60 animals per colony. If a breeder has several facilities producing the same stock, all stocks should be tested.

Lindsey, J. R. 1979. Historial foundations. In H. J. Baker, J. R. Lindsey, and S. H. Weisbroth, eds. The Laboratory Rat. Vol. 1. Biology and Diseases. Academic Press, New York.

Morse, H. C. III. 1981. The laboratory mouse—A historical perspective. In H. L. Fosters, J. D. Small, and J. G. Fox, eds. The Mouse in Biomedical Research. Vol. 1. History, Genetics, and Wild Mice. Academic Press, New York.

Nei, N. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.

Philip A. Wood

Professor, Department of Comparative Medicine

University of Alabama

Birmingham, Alabama

Genetic differences among animals can often lead to differences in phenotype. Now that we are in the era of creating animals with planned genetic differences, particularly "gene knockout" mice, we often anticipate what their phenotypes will be. Frequently, however, there may be no abnormalities or there may be unexpected abnormalities resulting from the intended genetic change. Frequently there are unexpected interactions with downstream pathways whereby the genetic background can markedly influence the phenotype resulting from a specific genetic change. There also have been gene knockout mice declared as having no abnormal phenotype; but when subsequent more specialized analyses were completed, striking abnormal phenotypes were discovered. Not only will genetic background significantly affect the phenotype of any given gene mutation as discussed by others at this meeting, but common environmental influences such as diet or cryptic infectious disease may also have a profound influence on the overall phenotype. The goal of this paper is to discuss a general approach for carefully assessing the many important influences on phenotype that are not often readily apparent at first glance.

It seems to me that the issue of phenotyping genetically altered animals is so complex, and subject to so many subtle factors within the animal as well as its environment, that we must begin thinking in terms of paradigms. I describe here a paradigm to consider when approaching phenotype assessment of mice and rats. This paradigm is offered as an approach undergoing further refining as our assessment tools improve. I have divided this systematic approach into primary and secondary levels of assessment for the simple reason that all possible analyses are not practical for any animals. Additionally, the primary level assessment

can and should be available for investigators at most biomedical research institutions, but the secondary level of assessment will likely require more specialized expertise and equipment and could be integrated into nationally based networks established for phenotype assessment. Both components will be crucial for fully assessing phenotypes and fully using the vast number of rodent models currently being developed and studied.

The goal of the primary assessment is simply to find abnormalities, through the following:

1.	Clinical Assessment: Many knockout mice are initially on C57BL/6 × 129/Sv hybrid background, therefore controls should be littermate controls with a similar mixed background.

a.	Litter size: number born/weaned, sex, and genotype distribution

b.  

Visual observation, particularly during the dark cycle when rodents are most active. Observe for behaviors that are aggressive, hyperactive, hypoactive, and so forth

c.  

Observe for any coat color differences, skeletal or other body conformational changes, and failure to thrive.

2.	Pathologic examination: Recommend evaluating both weanlings and retired breeders

a.	General necropsy to observe for any gross lesions and histopathology of all organs by an experienced rodent pathologist

b.  

Microbiologic/serologic/parasite evaluation to detect any background infectious disease that may confuse the phenotype resulting from a gene mutation.

c.  

Clinical pathology measures such as blood counts and simple urine analysis for protein and glucose.

d.  

Determine life span and reevaluate phenotypes in old age.

The goals of the secondary assessment are to evaluate and quantify the abnormalities found during the primary assessment. This will often require more specialized expertise and technology.

1.	Embryologic evaluation

a.	If abnormal litter size and genotype distribution are observed, these animals should be evaluated for gestational loss versus neonatal loss.

• This evaluation often requires timed matings with careful embryologic evaluation to detect the specific gestational stages when the animals die.

2.	Specialized pathologic evaluation

a.	Specialized stains for lesions detected by standard workup

b.  

Electron microscopy for cellular lesions that are not discernable at the light microscope level

c.  

Further evaluation of any blood cell count abnormalities with FACS analysis of leukocytes and other more specific immunologic measures

d.  

Specialized organ assessment such as specialty pathologic evaluation of heart changes, eye changes, bone changes, analyses of neuron/neurotransmitter distribution, and so forth.

3.	Specialized biochemical analyses

a.	Metabolite analyses on blood, urine, tissue extracts for specific metabolites such as amino acids, lipids, carbohydrates in deficient or excessive concentrations. These assays may require very specialized equipment and expertise with small sample size.

b.  

Enzyme or other specific protein analyses. This analysis would include not only assays that demonstrate the presence or absence of a protein, but also functional assays that may be crucial for corroborating any abnormal metabolite assays or blood cell abnormalities.

c.  

Hormone analyses. This analysis can be particularly important in diabetic animals as well as those with failure to thrive, small body size, infertility, skeletal abnormalities, behavior abnormalities, or skin disease.

4.	Physiologic assessment

a.

Pathologic evaluation may indicate organ dysfunction such as hyperplastic or hypertrophic enlargement, atrophy, or absence. Technologies are being developed to more thoroughly assess physiologic function such as miniaturized equipment that can transmit data via telemetry for these valuable physiologic measures in the awake unrestrained animal. Miniaturized instrumentation for procedures such as ultrasonography, magnetic imaging, DEXA analyses, indirect calorimetry studies, and other such devices are becoming increasingly available for these specialized measures in rodents, including those that are especially difficult in mice.

5.	Behavioral assessment

a.

This is an important, developing area of biomedical research that will take advantage of the numerous genetic modeling approaches provided by rodents. There already are many behavioral differences observed among the inbred strains of rodents. With the gene mapping tools now available many genotype/phenotype correlations can be pursued including studies pursuing the genetic components of drug abuse and mental illness. There are knockout mice that also have abnormal behavior that need evaluation. This will require not only the current behavior assess-