Project 1: The Mouse Brain Library Project 2: Internet Microscopy (iScope) Project 3: Neurocartographer and Segmentation of the MBL Project 4: The Neurogenetics Tool Box


















Principal Investigator/Program Director Williams,Robert W.

Core B (Genotyping and Mouse Colony Core) 

1. Objectives

The GMCC will provide essential resources that will enable the research community to greatly improve the speed and efficiency of mapping genes that modulate the size and cell composition of the mouse brain. We will generate close to 400,000 genotypes for a set of approximately 1400 mice (1200 animals between the ages of 46 and 76 days, and ~200 animals more than 2 years old) during this grant period. These genotypes will be a vital component of the Neurogenetics Tool Box. The brains of mice that have been genotyped by this core will be processed by the Neurohistology Core, and high-resolution images of all the cases will be added to the Mouse Brain Library (MBL). Genotype databases will be consolidated with the quantitative data generated as part of the MBL and the NeuroCartographer Project (Project 3). When this work is complete, the Neurogenetics Tool Box will contain a huge body of phenotypic and genotypic data for a population of over 1400 adult animals from a unique type of genetic cross. These data will make it possible to routinely map QTLs with a precision of <4 cM.

The GMCC will also be responsible for oversight of our mouse colony and for providing fixed brains to the Neurohistology Core. The maintenance of the mouse colony is not otherwise discussed in this Core description because these duties will be covered primarily by the UT Memphis Department of Comparative Medicine. Intramural funds provided by the Department of Anatomy and Neurobiology will cover minor expenses associated with perfusion and dissection of brains. We routinely process several thousand cases per year as part of other funded projects that deal with the structure of the mouse eye and retina. Additional supplies or technical assistance will not be required for the duration of this grant period.

Current status.

A large number of animals have already been generated for the principal cross that will be used in this program project. More than 1100 brains have been dissected and are now being prepared for embedding in celloidin. DNAs from 1152 of these animals have been extracted and arrayed in a set of 12 96-well microtiter plates. All these cases have been processed in the same way as the current set of 600 brains in the MBL. We also have breeding colonies for numerous standard and recombinant inbred strains and will therefore be able to supply the Neurohistology Core with the material required to have at least 12 animals per strain in the MBL.

A high-resolution mapping panel. 

We have generated what is called a tenth-generation advanced intercross. The cross started with two common but highly divergent inbred strains of mice: C57BL/6J (B for its black coat) and DBA/2J (D for its dilute brown coat). These are the same two parental strains used to produce the 36 BXD recombinant inbred strains that are already a major gene mapping resource in the MBL. These two strains differ genetically at more than 3000 genes and marker loci, greatly facilitating high-resolution QTL mapping. Of equal importance, the CNS architecture of these strains differs greatly (Williams et al. 1996, 1998, 2000); for example, weights of their brains are 420 mg and 500 mg, respectively. We have already made a high-resolution atlas of the brain of C57BL/6J (< and expect to do the same for DBA/2J. As part of the NeuroCartographer project, we will generate consensus 3D models of the brains of these two parental strains.

The tenth-generation BD intercross we are using has many advantages over a conventional intercross or backcross. With a conventional intercross or backcross, QTLs can generally be mapped only with a precision of 10 cM, even when 2000 or more cases are studied. To overcome this problem, Darvasi (1998) suggested the advanced intercross as a conceptually simple method to increase the precision of QTL mapping to 2 cM. An advanced intercross accumulates recombination events over many generations (see below), effectively stretching the genetic map. The genetic map doubles in length with each doubling of generation number. For example, a G4 cross has twice the map length of an F2 cross. The genetic map of G10 animals is five times as long as the map based on a conventional cross7000 cM rather than 1400 cM. This means that for a given level of effort, the G10 panel provides five times the precision of a F2 intercross. Thus, positional candidate gene cloning approaches become far more feasible. Dr. K. Manly has recoded the Map Manger QT program to create Map Manager QTX, which will accommodate data sets from advanced intercrosses (Manly and Olson 1999). Project 4, the Neurogenetics Tool Box, will incorporate both genotypes and quantitative neuroanatomical traits from the entire G10 panel.

Advantages for Neuroinformatics. 

Perhaps the most important advantage of the advanced intercross for neuroscientists is that large numbers of different morphometric traits will be studied in a single set of animals. This means that we will be able to explore the complex network of gene interactions that are undoubtedly important in CNS development and aging. To make this idea more concrete, consider the possibility that a QTL specifically controls variation in the proliferation of granule cells in three very different parts of the CNSthe olfactory bulb, the dentate gyrus, and the cerebellum. A single researcher might have the time and energy to study one of these populations, but not all three. The link among all three regions would remain unsuspected and undetected. But by providing a large community resource, the Genotyping and Mouse Colony Core, in concert with the Neurogenetics Tool Box, makes all information cumulative in a single experimental cross. If three investigators in succession, studying three different CNS regions, map what might be called granule cell number modulate genes to the same chromosomal interval, then the genetic link to a particular cell type across regions will become obvious with a look at the list of mapped quantitative trait loci. Thus, a community resource can expose genetic correlations (pleiotropy) among connected and unconnected parts of the brain.

Statistical Power. 

Virtually all of the CNS traits that we anticipate mapping are complex multigenic traits. The question arises as to what statistical power the G10 progeny set will have to detect QTLs. How small a QTL effect will this panel be able to detect? Darvasi (1998) provides an equation for the approximate number of progeny for a power of 0.5 using an advanced intercross at 75/(2d2 + h2) where d is the standardized additive allele effect size and h is the standardized dominance effect size. For a locus with a purely additive effect of 0.20 standard deviation, 934 animals are needed for a power of 0.5. Thus with our G10 sample we anticipate achieving a power of >0.5 for QTLs that have a standardized effect of only 0.2 SD. In other words, this is a remarkably sensitive cross for detecting and mapping QTLs.

2. Staffing and Oversight

The GMCC will be staffed by one full-time professional. Dr. Jing Gu will be responsible for the extensive genotyping. Dr. Gu currently runs the PIs genotyping lab. Dr. Gu will also be responsible for error checking and for entry of genotype data into our consolidated relational databases. She will be supported by the PPG and will be responsible for genotyping the G10 advanced intercross. She is highly skilled in all aspects of genotyping. The Molecular Resources Center will provide additional personnel to assist with the work.

3. Resources and Environment

The Genotyping and Mouse Colony Core will be housed at the University of Tennessee, Memphis. This Core will initially share facilities in the PIs lab (311 Wittenborg). However, late in Year 01 we will open a new UT Memphis Genotyping Center as a division of the UT Memphis Molecular Resources Center. The Genotyping Center will be directed by Dr. Robert Williams. This facility will have an initial capacity of at least 200,000 genotypes a year, of which half will be devoted to this PPG (see letter from Dr. Michael Dockter, Vice Dean for Research).

4. Administration

Dr. Jing Gu will report to Drs. Williams and Manly. Drs. Williams and Gu will meet at least once a week to review progress and will keep closely in touch with Dr. Manly by email and phone. Dr. Gu will maintain several key databases that will be used by Dr. Manlys group as part of Project  4. Dr. Manly will have direct access to the FileMaker Pro server that will host the genotype databases.

Funds for the Genotyping and Mouse Colony Core will be managed by the Department of Anatomy and Neurobiology and by the Administrative Core of the Program Project. Funds to buy equipment to support the expanded genotyping throughput will be provided by the University of Tennessee. The limited support required to fix animals and ship brains to Dr. Rosen will be covered by intramural funds. The budget for this core does not include any equipment request. Our supplies request is limited to genotyping (~$0.20/genotype) and is relatively modest. Our cost is far lower than current commercial genotyping prices. For example, Research Genetics currently charges in excess of $4 per microsatellite genotype. We will be able to generate 400,000 genotypes for well under $80,000 in supplies.

5. Justification

Full-scale genotyping and extension of the Mouse Brain Library, as proposed in this application, will allow multiple research groups to study many traits in the same set of animals, maximizing the utility of each cross, preventing duplication of effort, and revealing the complex genetic basis of variation in different CNS compartments. This community effort also allows us to substantially increase the positional precision of QTLs that are mapped. The proposed project will greatly reduce the cost and effort of examining the basis of normal variation in CNS architecture among and within strains of mice.

Until now, it has been possible to adequately analyze only a small number of phenotypes in each animal that is genotyped. Each research group generated a unique F2 intercross or backcross consisting of several hundred animals. Researchers observed a small number of traits, genotyped their modest sample of animals, and subsequently mapped the loci modulating the observed traits. R. Williams and colleagues have used this approach to map more than a dozen QTLs that control morphometric and quantitative variation in the architecture of the eye and brain of normal strains of mice (see Appendix; Williams et al. 1998; Zhou and Williams 1999; Williams 2000). Only four trait values were acquired for most mice: body weight, brain weight, eye weight, and retinal ganglion cell number. Even this was a huge undertaking for a single lab. This cottage industry approach is incredibly inefficient and yields only low-resolution estimates of the chromosomal location of QTLs and their phenotypic effects.

Use of the core by individual projects. 

The Genotyping and Mouse Colony Core is a key adjunct of the Neurogenetics Tool Box, and all of the genotype data will be directly integrated by databases and programs described in Project 4.  This core will also provide brains to the Neurohistology Core for processing and entry into the MBL. The genotypes are far more useful when combined with the extensive phenotype data that will be generated in Project 1 (images), Project 2 (cytological assays using the iScope), and Project 3 (quantitative segmentation of brains). The genotype data are a permanent tool that can be used to genetically dissect any quantitative (or qualitative) trait exposed by data sets in the MBL or exposed by the iScope or NeuroCartographer projects.

6. Procedures

Generating the AI cross. 

Our G10 advanced intercross was generated using the method outlined in Darvasi (1998) with modifications suggested by Dr. Lee Silver. We first generated (C57BL/6J x DBA/2J) F1 mice and intercrossed them to generate a standard B6D2 F2 generation. We also carried out the reciprocal cross to generate D2B6 F2 animals. Approximately 100 of each of these reciprocal F2 types were generated. Thirty mating cages were set up, and one or two F2 individuals were selected from each litter and mated to generate the G3 progeny. Siblings were never placed together in a mating cage; since matings were nonfilial, we use a G prefix rather than the F prefix for the third generation. The G3 offspring from different cages were mated to generate G4. This procedure was continued until G10 with a constant attempt to minimize potential fixation/inbreeding. Both sexes have been collected, and the range of ages of the current set is quite narrow (between 46 and 76 days). Two hundred animals have been set aside as part of an aging colony. They will be sacrificed at about 700 days of age. The breeding history of the advanced intercross is currently maintained in a FileMaker Pro relational database, and a large part of the database (through August 1998) is available online at <>. The entire database will soon be published online with full genotypes.

Scanning the genome for QTLs. 

To perform a genome-wide scan of the G10 with an average separation between markers of 20 cM, we will employ 350400 polymorphic microsatellite markers. To choose PCR primers for whole-genome QTL scanning, we select 300400 evenly spaced marker loci in which the product differences between parental strains are greater than 2 base pairs. Selection is not difficult because these strains have already been typed at many hundreds of microsatellite loci (Dietrich et al. 1994). The G10 genetic map is about 7000 cM long, so the average distance between any locus and its neighboring microsatellite loci is 18 cM. Typing 1000 G10 progeny with each of 350400 markers requires 350,000400,000 PCR reactions, which can be accomplished over 3 to 4 years (mean daily output for one technician is currently 384 reactions). We are investigating procedures that may allow us to double this output.

Genotyping the AI cross. 

The PCR reaction (10 l) consists of 40 ng of genomic DNA extracted from spleen or tail using an inexpensive high-salt procedure, 1 M of each primer, 1x reaction buffer, 200 M dNTPs, 2.5 mM MgCl2, and 0.5 U of Taq polymerase. Reactions are carried out in a thermocycler (M.J. Research) with a hot bonnet using a touch-down procedure in the first 5 cycles with a high annealing temperature to improve specificity. The PCR products are electrophoretically separated on either agarose (TMC MethPhor, 3.5%) or acrylamide gels and visualized with ethidium bromide under UV illumination. Agarose is preferred when alleles differ by more than 8 base pairs, acrylamide when alleles differ by 28 bp. We have considerable experience in the use and selection of microsatellites for typing alleles in test crosses and in running these reactions and gels efficiently. Data will be entered into a relational database (FileMaker Pro) and exported to the Neurogenetics Tool Box for statistical analysis and QTL mapping.

Data entry. 

All genotypes will be entered into a FileMaker Pro database. For simplicity in QTL mapping and multiple regression, markers with the BB homozygous genotype will be scored as +1, the BD heterozygote as 0, and the DD homozygote 1. A failed reaction will be scored as a non-numerical value (U). We will not score any markers with dominant molecular banding patterns. Apparent double recombinants which are defined by a single discordant typing will be retested to reduce typing errors. Animal ID number will be the key field in relating genotypes with MBL phenotype databases. R. Williams and K. Manly will design all genotype databases. Dr. Manly will use these databases in the Neurogenetics Tool Box project. Dr. Manly will have direct access to genotypes generated at UT Memphis using a web interface to our FileMaker Pro database (see <> for examples of these types of web databases).


Many additional mice will be incorporated in the MBL collection over the period of this grant. Dr. Rosen, PI of the Neurohistology Core and the MBL, will order mice from our Administrative Core assistant, who will obtain them from the Jackson Laboratory. Mice will be anesthetized with Avertin (0.5 to 0.8 ml, i.p.) and perfused transcardially with  0.9% phosphate buffered saline followed by approximately 15 ml of 1.25% glutaraldehyde and 1.0% paraformaldehyde in 0.1 M phosphate buffer. An additional 10 to 20 ml of double-strength fixative (2.5% glutaraldehyde and 2.0% paraformaldehyde in the same buffer) is subsequently injected for 1 to 2 minutes at an increased rate and pressure. The head is removed and put in the final fixative until the brain is dissected and weighed. Brains in fixative will be shipped in groups of 50 to 100 to Dr. Rosen. We have followed this procedure without any problems for the past two years.

7. Financial Considerations

Technical support. A full-time senior technician is needed to carry out activities that are part of this Core.

Equipment. The Vice Dean of Research at UT Memphis will provide $140,000 for equipment required to perform as many as 400,000 genotypes over the duration of this grant. The PI will use these funds to assemble a new high-throughput genotyping lab. The appended letter from the Dean lists some of the key equipment that will be purchased.

Travel costs. None

8. Bibliography (Core B)

Darvasi A, Soller M (1995) Advanced intercross lines, an experimental population for fine genetic mapping. Genetics 141:11991207.

Darvasi A (1998) Experimental strategies for the genetic dissection of complex traits in animal models. Nat Genet 18:1924.

Dietrich WF, Miller JC, Steen RG, et al. (1994) A genetic map of the 4006 simple sequence length polymorphisms. Nat Genet 7:220245Doerge RW, Churchill GA (1996) Permutation tests for multiple loci affecting a quantitative character. Genetics 142:285291.

Manly, K.F. and Olson, J.M.  (1999) Overview of QTL mapping software and introduction to Map Manager QT. Mamm Gen 10: 327-334.

Roff DA (1997) Evolutionary quantitative genetics. Chapman Hall, New York.

Schmitz S., Cherny SS, Fulker DW (1998) Increase in power through multivariate analyses. Behav Genet 28:357363.

Williams RW (1998) Neuroscience meets quantitative genetics: Using morphometric data to map genes that modulate CNS architecture. In: Morrison J, Hof P (eds) Short course in quantitative neuroanatomy. Society of Neuroscience, Washington DC, pp 6678.

Williams RW (2000) Mapping genes that modulate mouse brain development: A quantitative genetic approach. In: Goffinet A, Rakic P (eds) Mouse Brain Development. Springer, Berlin (2000) in press.

Zhou G, Williams RW (1999) Eye1 and Eye2: Gene loci that modulate eye size, lens weight, and retinal area in mouse. Invest Ophthalmol Vis Sci 40:817825.


e. Human Subjects none

f.  Vertebrate Animals

Description of proposed use of animals

i.      We will use the species Mus domesticus in these studies. Most animals are originally obtained from the Jackson Laboratory, Bar Harbor, Maine. Both sexes will be used. We will have breeding colonies and intercrosses of selected strains. We will have a colony of aged mice to study the genetics of CNS aging.

ii.     Justification of animal use. Mice will be used to characterize the genetics of CNS. Over the period of this grant, we anticipate that we will use an average of approximately 1000 mice per year. Such large numbers of animals are required in order to map quantitative trait loci with precision. All of these animals will be generated in our mouse colony.

iii.     Veterinary care. The mice will be maintained in departmental AAALAC approved facilities under the supervision of the Department of Comparative Medicine of the University of Tennessee. Only trained personnel will work with our mouse colony, and they will check room conditions, cage, and animal status daily. The Department of Comparative Medicine has new facilities in the Nash Annex. This is a state-of-the-art vivarium with extensive space for pathogen-free colonies of mice. All of our colony will be kept in this facility. The facility employs two full-time veterinarians to monitor the health and well-being of animals.

iv.    Analgesics and anesthetics. NA

v.     Euthanasia method. Adult mice that will be fixed by transcardial perfusion will be anesthetized deeply with Avertin. Adult mice that will be used without perfusion will be sacrificed by cervical dislocation. Neonates and fetal animals are decapitated with scissors. Some mice will be euthanized by exposure to 100% CO2. These methods are consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association.

i. Consultants Dr. Warren Young (Scripps Institute) will serve as a consultant on Projects 1 .

h. Consortium Arrangements see attached documents and letters of intent