The explosion in the technology of molecular genetics gives unprecedented opportunities to explore the genetics of many species. The technologies are already being applied in dairy cattle research and may well affect the way we breed cattle in the future.

In Part 1, the genotype of an animal was compared to a manual of instructions consisting of eleven hundred (1100) volumes each of 1000 pages written in a language we barely understand. The critical information is carried by approximately one hundred thousand genes. Each gene corresponds to one page of instructions but for every page of useful instructions there are also ten pages of nonsense. Every animal carries two complete copies of the manual but there are many variations in each page of instructions so that no two copies of the manual are ever identical. If the prospect of understanding this complex system is not daunting enough, perhaps the most remarkable thing is that each copy of the manual is stored in a molecule of DNA which weighs a mere one million millionth of a gram.

The basic alphabet used in the genetic instruction manual, the genotype, was deciphered in the 1950's, but it is only within the past few years that tools have been developed for reading the instruction manual of animals. Even so, the process of reading is slow and expensive. The first complex genotype to be read completely will be our own, under the internationally co-ordinated Human Genome Project. This mammoth task will probably be completed within the next 5 years, at a cost of several billion dollars, involving over 10,000 scientists. Knowing the string of three to four billion letters that makes up the instruction manual is an impressive achievement, but it is a very long way from being able to understand the manual, since we still know very little about the language. If this seems strange when we know the alphabet, imagine how little use it would be knowing that French and English use essentially the same alphabet, if you tried to read a book in French without having learned the French language (or vice versa). Yet, this would be a trivial task compared to understanding the genetic language, since French and English are very similar to each other, whereas the genetic language is totally different to any other language we know.

Although current research in molecular genetics is progressing at an incredible rate, it is clear that full understanding of the genetic language is a long way off. Dairy cattle are far more expensive to work with than laboratory animals, and are not considered as important as humans, and so only a tiny fraction of the billions of dollars spent in this area will come to dairy cattle. But, the genotype of all animals share a great deal in common, so that much of the information gained from other species will ultimately be useful for dairy cattle research. In the meantime, there are molecular technologies that are already of potential interest in dairy cattle breeding.

A variety of technologies developed through the 1970's opened the way to identify specific sequences in an animal's genotype. This is akin to searching a manual for a particular word, sentence or page of information. This would be an impossible task to do by hand, but just as a computer can now search a national data base and rapidly find the records for one particular cow in a particular farm, so molecular techniques can search through the vast genotype and find a particular sequence. These sequences may be part of a particular gene of known function, equivalent to a particular page in the manual, where we know the basic product of that page's instructions. The sequences may, however, be anonymous, a page or chapter number in the manual, where we know nothing about the instructions in that area of the manual. Both types of sequence can be useful.

Although some of the technologies for finding sequences have been around for 15 years or more, it is the related and very recent technology of polymerase chain reaction (PCR) that has made them accessible on a routine basis. We noted earlier that the genotype is encoded in the DNA, a single molecule of which weighs only one million millionth of a gram. Extracting sufficient DNA to carry out genetic tests required large samples of tissue, blood or semen cells. PCR is a chemical reaction which allows a very small amount of DNA, even a single molecule, to be copied many millions of times in just a few hours. It is a chain reaction, where a single molecule is copied to make 2 copies, these 2 are copied to make 4 copies, the 4 become 8, the 8 become 16 and so on, until a minute quantity of DNA has been multiplied into a quantity sufficient to perform genetic tests with relative ease.

Several genes that can be identified using a molecular test are already familiar to dairy cattle breeders. K-casein was much in the news a few years ago. The k-casein gene is a set of instructions coding for the production of the k-casein protein, which forms only a small proportion of total milk protein, but has an important role in milk processing properties. There are two principal variants of k-casein, the A and B variants. Cows carrying the B variant have been promoted as having higher protein yields and better cheese making properties. Chemical tests of milk have been available for many years, and a molecular test was available in the mid-1980's, but PCR allowed development of a rapid test that could be applied with very small samples of blood or milk of cows, and blood or semen of bulls. Most bulls are now routinely tested for their k-casein status (since all animals have two copies of the genetic instruction manual, the possible genotypes are AA, AB or BB). Routine testing of cows has not caught on because the original claims for the advantages of the B type have faded somewhat, and it clearly does not pay to test cows.

Another example of a recently identified gene is the BLAD gene (bovine leucocyte adhesion deficiency), where animals with two copies of a defective version suffer severe illness and die young. Young bulls are now tested prior to entering AI, and most carriers (those that carry one copy of the defective version) are eliminated so that BLAD should soon become a disease of the past.

An example of molecular tests not involving a specific gene are the sexing probes. These tests identify a specific sequence that is carried only by males. This sequence is not itself a set of instructions; it appears to be part of the large amount of non-coding padding material that the genetic manual carries. It does, however, uniquely identify the version of the chapter in the genetic manual that contains instructions for being male from the version that contains instructions for being female. It is like having two alternative chapter headings, say "Chapter 426. Instructions Determining Sex: Male Version" versus "Chapter 426. Instructions Determining Sex: Female Version". Using PCR, an embryo can be tested by removing just one cell and only the embryos of the required sex need be transferred. This is already available as a routine service in several laboratories. It is still a fairly costly procedure, currently limiting its use to elite breeders.

Similar procedures identify a huge class of variable non-coding sequences called micro-satellite markers. They are not instructions themselves, but mark versions of different regions in the genome; again they are analogous to identifiers for different versions of different chapters. Even without knowing anything about the contents of these chapters, or even where they appear in the manual, these markers may be useful to us. The following is one of several possible strategies that might be used.

Every sire carries two copies of the genetic instruction manual. He passes one copy on to each of his progeny and this copy is made up of sections taken at random from each of the two copies he carries. Imagine the sire carries two versions of a chapter in the genetic manual, say Chapters 355a and 355b. He has a number of sons entered into AI, half of which carry his version 355a and half carry his 355b. If version 355a contains a better set of instructions that leads to higher milk yield than version 355b, we will see this as a difference in the average proof of the two groups of sons. Having identified this difference we can then pre-select young bulls before entering AI, on the basis of which version of Chapter 355 they inherit from their sire. This should improve the average merit of the young bulls being tested and may allow us to reduce the number of bulls we need to test, hence reducing costs.

The beauty of this approach is that we do not need to know anything about the contents of a particular section of the manual, as long as we can trace with our markers which version of a particular section is inherited from a particular sire. The disadvantage is that we do not know in advance which sections have versions for enhanced performance, nor which sires carry which versions. To identify useful markers will involve studying several hundred different markers in all available sire families. Despite recent advances in technology this remains a sizeable task and will require the co-operation of many research groups.

The AI studs in Canada are currently co-operating to create a Canadian Bovine DNA Bank, which will hold DNA from all young bulls being tested now and in the future, as well as DNA from all past and present bulls for which semen samples are still available. This DNA bank will be held by the Saskatchewan Research Council Blood Typing Laboratory and samples will be made available to Canadian researchers with projects studying molecular genetic inheritance in cattle.

The Canadian Bovine DNA projects will put us in a position to utilize discoveries as they are made and help keep Canada at the forefront of dairy cattle breeding technology. It is still too early to say how much impact molecular technologies will have on dairy cattle breeding. It is unlikely that we will understand sufficient about the genome for molecular information to replace conventional genetic evaluation procedures such as progeny testing. But, we have already seen a few molecular tests made available, and we can expect to see an increasing number of tests in the future.