A breeding mouse colony differs significantly from a static one in the type and complexity of information that is generated. In a non-breeding colony, there are only the animals and the results obtained from observations and experiments on each one. In a breeding colony, there are animals, matings, and litters, with specific genetic connections among various members of each of these data sets. Classical genetic analysis is based on the transmission of information between generations, and as a consequence, the network of associations among individual components of a colony is as important as the components are in and of themselves.
An ideal record-keeping system would allow one to keep track of: (1) individual animals, their ancestors, siblings and descendants; (2) matings between animals; (3) litters born from such matings, and the individuals within litters that are used in experiments or to set up the next generation of matings; and (4) experimental material (tissues and DNA samples) obtained from individual animals. Ideally, one would like to maintain records in a format that readily allows one to determine the relationship, if any, that exists between any two or more components of the colony, past or present.
Based on these general requirements, two different systems for record-keeping have been developed by mouse geneticists over the last 60 years. The "mating-unit" system focuses on the mating pair as the primary unit for record-keeping. The "animal/litter" system treats each animal and litter as a separate entity. As discussed below, there are advantages and disadvantages to each approach.
With this system, each mating unit is assigned a unique number and is given an individual record. When record-keeping is carried out with a notebook and pencil, each mating pair is assigned a page in the book. The cage that holds the mating pair can be identified with a simple card on which the record number is indicated; this provides immediate access to the corresponding page in the record book.
When litters are born, they are recorded within the mating record. Each litter is normally given one line on which the following information is recorded in defined columns: (1) a number indicating whether it is the first, second, third, or a subsequent litter born to the particular mating pair; (2) the date of birth; (3) the number of pups; and (4) other characteristics of the litter that are of importance to the investigator. Individual mice within any litter can be identified uniquely with a code that includes the mating unit number, followed by a hyphen, the litter number, and a letter that distinguishes siblings from each other. For example, the fourth pup in the third litter born to mating unit 7371 would be numbered 7371-3d. This system provides for the individual numbering of animals in a manner that immediately allows one to identify siblings and parents.
At the outset, parental numbers are incorporated into each mating record, and since these are linked implicitly to the litters from which they come, it becomes possible to trace a complete pedigree back from any individual. It also becomes possible to trace pedigrees forward if, as a matter of course, one cross-references all new matings within the litter records from which the parents derive. For example, if one sets up a new mating unit that is assigned the number 8765 with female 5678-2e and male 5543-1c, the number 8765 could be inscribed on appropriate lines in records 5678 and 5543. 21
There are several important advantages to a record-keeping system based on the mating unit: (1) only a single set of primary record numbers is required; (2) one can easily keep track of the reproductive history of each mating pair; and (3) information on siblings can be readily viewed within a single location. Furthermore, it is easy to incorporate this system of record-keeping into a simple spreadsheet file that can be maintained on a desktop computer or fileserver. This can be accomplished most readily by having each row represent an individual litter (or even individuals within a litter) with columns for (1) Mating unit number, (2) Father's number, (3) Mother's number, (4) Litter number, (5) Birth date, (6) Number of pups, and (7) further information. To record information on individuals within particular litters, new rows having the same format can be formed but with the litter number-sibling letter combination used in place of the litter number alone in column 4.
New litters can be recorded initially as they are born in empty rows at the bottom of the file. By re-sorting the database according to the first column, an investigator would be able to see all litters born to a particular mating pair in sequential rows. Upon sorting according to birth date, the list of litters could be displayed according to age. With search or find commands, it would be possible to identify ancestors, descendants, and siblings related to each animal.
The major disadvantage to this form of record-keeping is that records are focused on mating pairs and litters rather than individual animals. Thus, it is not well-suited for investigators who need to record and retrieve animal-specific information. It is also less than ideal for situations where the mating unit is not sacrosanct and animals are frequently moved from one mate to another. Under these circumstances, the animal/litter system described below is more appropriate for record-keeping.
In a second system developed originally by one of the earliest mouse geneticists, L. C. Dunn, there are two primary units for record-keeping the individual animal and the individual litter. Each breeding animal is assigned a unique sequential number (at the time of weaning) that is associated with an individual record occupying one row across facing pages within an "animal record book" or a spreadsheet file. Each animal record contains the numbers of both parents and through these it is possible to trace back pedigrees. A separate "litter record book" or spreadsheet file is used to keep track of litters which are also assigned unique sequential numbers attached to one row in the database. In this record-keeping system, animal numbers and litter numbers are assigned independently of each other.
A third independent set of numbers are those assigned to individual cages. Cage numbers can be assigned in a systematic manner so that related matings are in cages with related numbers. For example, different matings that derive from the same founder of a particular transgenic line may be placed in cages numbered from 2311 to 2319. A second set of matings that carry the same transgene from a different founder could be placed into cages numbered 2321 to 2329, and so on. Thus, the cages between 2300 and 2399 would all have animals that carried the same transgene, however, different sets of ten would be used for different founder lines. For matings of animals with a second transgene, you might choose to use the cages numbered 2400 to 2499. This type of numbering allows one to classify cages which represent matings in a hierarchical manner. Although at any point in time, every cage in the colony will have a different number, once a particular cage is dismantled, its number can be re-assigned to a new mating. Cage cards from dismantled matings can be saved for future reference.
When a litter is born, the litter record is initiated with entries into a series of columns for (1) an identifying number, (2) the birth date, (3 & 4) the numbers of the parents, (5) the number of the cage in which the litter was born, (6) the number of pups born, and (7) any other information of importance to the investigator. In addition, the litter number can be inscribed on the cage card (which may or may not have additional information about the mating pair). When an animal is weaned from a litter for participation in the breeding program, an animal record is initiated. The most important information in the animal record is the number of the litter from which it came, the cage that it goes into, and the date of that move (all entered into pre-defined columns). The cage number is particularly important in allowing one to trace pedigrees forward from any individual at a future date. If an animal is moved from one cage to another at some later date, this can be added to the record in another column.
With three unrelated sets of numbers and the need for extensive cross-referencing, the animal/litter system is complex, and implementation on paper is labor-intensive. However, it does provide the investigator with additional power for analysis. For example, by choosing cage numbers wisely and saving cage cards in numerical order, it becomes possible to go back at any point in the future and look at all of the litters born to a particular category of matings over any period of time. With the mating unit system, this could only be accomplished by using different files for different categories of matings. However, it then becomes very difficult to keep track of matings formed with animals taken from different files.
Another difference between the mating unit system and the animal/litter system is the ease with which it is possible to keep track of animals that are moved from one mating unit to another. The mating unit system is most effective for colonies where "animals are mated for life." The animal/litter system is effective for colonies of this type as well, but is also amenable to those where animals are frequently switched from one mate to another.
The animal/litter system of record-keeping has been incorporated into a more extensive computer software package that greatly facilitates data entry, with automatic cross-referencing and extensive error checking. This software package, called the "Animal House Manager" or AMAN and MacAMAN, are specialized database programs that allow users to record and retrieve information on animals, litters, cages, tissue and DNA samples, and restriction digests generated from a breeding mouse colony (Silver, 1993b). Data are entered through a series of queries and answers (for AMAN) or pointing and clicking (for MacAMAN). With automatic cross-referencing and specialized protocols, the same information never has to be entered more than once. Hard-copy printouts can be obtained for cage cards, individual records, or sets of records (in abbreviated form) that have been recovered through searches for positive or negative matches to particular words or parameters. Search protocols are highly versatile; for example, it is possible to print out a cage-ordered list of litters that are old enough for weaning on a particular date, or a list of live mice from a particular set of breeding cages that are ordered according to the cage in which they were born. AMAN and MacAMAN provide investigators with the ability to maintain control over a complex breeding program with instant access to each record, current and past. It can store up to 100,000 records in each of four files for animals, litters, DNA/tissue samples, and restriction digests.
For licensing and other information, please refer to Appendix B.