Graph of genetic variation graph

Variety is the spice of life and the best tool in a breeder’s toolbelt

By Nicola Dennis

There is a whole stack of math that goes into an effective selection program. However, at its very basic level, genetic selection relies on having a variety of animals and estimating which ones have the best genes to pass on to their offspring.

If you can correctly choose the good parents (and particularly the sires) each year, then you can expect a year-on-year improvement on the proportion of “good” offspring that are born. But, what if you are not getting good year-on-year improvement? Let’s breakdown what could be going wrong.


1. Non-genetic variation got in the way

When estimating how good the sire was, some environmental or management influence may have crept which gave the sire an “extra edge” that is not seen in his progeny when they are farmed under other conditions. This can be because animals from different farms were not compared on a level playing field.

Genetic evaluation systems such as SIL can account for environmental differences if a sire has progeny raised on multiple farms. Non-genetic variation can also interfere through a genotype-by-environmental interaction (GxE) where some genotypes do better in certain environments but poorly in others.

Progeny test schemes such as the Beef + Lamb Genetics Central Progeny Test (CPT) which compares sire lines with experimental rigour in both hill and lowline environments are very important for mitigating this problem.

2. Reached the biological brick wall

Let us digress into thoroughbred racing for a moment. In the 1900’s the average winner of the Kentucky Derby took 130 seconds to complete the track (which means they were travelling about 55.7km/h). In the 1940’s, after another 40 years of genetic selection and better management, the winners were 3km/hr faster (about 58km/hr).

If that trend had continued, then we would expect the current Kentucky derby winners to be (after over 70 years of improvement) travelling at an average of 65 km/hr. As it happens the Kentucky Derby track record is 119.4 seconds or 60.7km/h (set in 1973) and there have been no real improvements since then.

There are biological limits to what an animal can achieve, at some point the animal lacks the basic machinery (wheels, perhaps?) to travel any faster. It seems that Thoroughbred breeders may have hit that limit. Some may also argue the problem with Thoroughbreds is that the studbook has been closed for 300 years, so the population is inbred and lacks genetic diversity. This leads us nicely onto the next potential problem.

3. The trait does not have enough genetic variation

Genetic variation can be too low to be effective for two reasons. Firstly, the population may be too small, either because there are too few animals around, or because recording is patchy. Animals at the extreme ends of the bell curve (either very good or very bad for the trait) are less frequent than average animals. Each animal that is recorded is like having a lotto ticket: the more tickets you hold the greater your chance of hitting the animal lottery and finding that extremely good animal that provides a huge improvement in the trait. If the population is small you may have to accept that a lot of times the “best” sire is not that much better than what you already had.

The second reason that you may find a lack of genetic variation is because there is no genetic variation. Some genes have very important and specific roles and variation in these genes usually means that the animal is not viable. Therefore, we do not see much variation between animals in these parts of the genome. An extreme example of a trait with no genetic variation for this reason would be the number of heads on a sheep; it goes without saying that the optimum number is “1”. Animals with more or fewer heads are born but do not survive in the general population. 


Assuming that we have a good population size and are not selecting for things that are detrimental to the survival of the animal, do we need to worry about losing genetic variation when we do selection? Do the things we select for now hurt the amount of genetic variation we will have in the future?

According to the experts (and here I should credit Teagasc Geneticist Donagh Berry for explaining this so well at a recent conference), if you had genetic variation, the evidence suggests you can’t really “lose it” even under extreme selection or high levels of inbreeding. Those inbred Kentucky-derby-winners that are travelling maximum speed still breed a considerable number progeny that can’t outrun a tin of pet food because the genetic variation for speed still exists.

One of the reasons that we can’t really “get rid of” genetic variation is because it is an ongoing phenomenon. Genes mutate at a constant rate which means that each generation picks up a handful of mutations that were not present in their parents.

There is a fantastic example in plants. For over a century, the University of Illinois has been experimenting on a single genetic line of maize (The Illinois Long-Term Selection Experiment established in 1896). From the original plant, they have bred a line that is extremely high in oil content and a line that is extremely low in oil content. When the low oil line seemed to get down to minimum oil content around the 1960’s (the high oil line continues to show genetic gains to this day), they decided to see what would happen if they switched the lines around.

Could the extreme low line produce high oil content with selection? Yes, it could!

When selection was reversed, the low line showed the same genetic trend as the line that had been high the whole time (albeit about 50 years behind). The high line managed to get back down to minimum oil levels in the same amount of time as the original low line. The tools to make the switch (genetic variation) were still available in the population even after 50 years of extreme selection. It just goes to show that it is never too late to change your breeding goals!