Summary: Knowledge of which traits are antagonistic can be utilized to manage the impact of selection decisions on other correlated traits. However, it is important to remember that although genetic correlations can sometimes create the need to exercise more care in selection to alleviate unintended consequences, these correlations can sometimes be utilized to our benefit. Understanding the magnitude and direction of genetic correlations can assist in selection decisions. Utilizing balanced selection for multiple EPDs in a breeding objective or using an appropriate selection index will ensure that genetic antagonisms don’t become a limiting factor for genetic progress.
It is important to remember that selection decisions in livestock often impact other traits which may not otherwise be under selection. These unintended consequences are a result of relationships that exist between traits, which can be described by a genetic correlation.
A genetic correlation provides information on whether two heritable traits share genes and they range from 1 to -1. In reality, correlations are rarely 1 or -1. They are generally somewhere between these values. If two traits are not correlated, they are controlled by different genes within the genome and their genetic correlation is zero (Figure 1, panel A). Uncorrelated traits should be completely independent and selecting for increased genetic merit in one trait should have no impact on the other trait. If they are correlated, it means that selection on one trait will cause changes in the other. If one considers traits that are genetically correlated, such as weaning weight and yearling weight, this concept is fairly straightforward, because genes that control growth at one stage of an animal’s life would be logical candidates to impact growth at a later stage in their lives. The genetic correlation could be very large (closer to 1 or -1 than 0) as is the case between growth traits, with most genes being shared between the two traits with fewer genes that can be selected for independently (Figure 1, panel C), or very small (near zero), where few genes are shared between the traits and the majority of the genes that impact trait one are independent of trait two (Figure 1, panel B). When selecting for a trait that is highly correlated to another trait, we can expect substantial changes in the other trait simply due to the strong genetic correlation between the trait we are selecting on and the trait to which it is correlated. For two traits with a low genetic correlation, selection on one of the traits has an impact on the other, but to a lesser degree.
Genetic correlations are classified by the strength of the relationship (low from 0 to ± 0.2, moderate from > ± 0.2 to ± 0.6, and high > ± 0.6 to ± 1.0) between two traits and its directionality. These two things, strength and directionality, determine if a genetic correlation is advantageous or not. Genetic correlations can be either positive or negative, which reflects the direction of the relationship between the two traits. Genetic correlations can also be classified as to whether one finds their relationship desirable or not desirable.
A positive genetic correlation (Figure 2) simply means that as one trait increases, the other trait also tends to increase. Even though the word positive tends to provide a favorable connotation, a positive genetic correlation doesn’t necessarily mean that the relationship is favorable. A negative genetic correlation (Figure 2) indicates that as one trait increases, the other trait tends to decrease. Negative seems to imply that the relationship is not advantageous, but that is not necessarily the case. It only tells us the direction of that relationship. A genetic correlation is favorable when selection on one trait produces a desirable outcome in another trait. As a result, a trait can have a positive favorable correlation, or a negative favorable correlation. Traits that are cheap or easy to measure but that have favorable genetic correlations with economically important traits that are more difficult or expensive to measure can be utilized as indicator traits. Traits can also have unfavorable genetic correlations. Unfavorable genetic correlations are sometimes referred to as genetic antagonisms. Genetic antagonisms cause decreases in genetic merit for some traits when single-trait selection is practiced or when failing to consider selection responses in correlated traits that are not directly under selection.
Examples of Genetically Correlated Traits
Genetic correlations between traits are fairly common. Table 1 lists some traits that are genetically correlated. An example of two traits that are positively correlated include growth traits, such as weaning weight and yearling weight. As selection is practiced to increase weaning weight, yearling weight tends to also increase even though selection is not being directly practiced on yearling weight. Growth traits also encompasses birth weight, which has a positive yet antagonistic relationship with weaning weight. When selecting for larger weaning weights, there tends to be an increase in birth weight if selection is not practiced for both traits simultaneously. Birth weight is also genetically correlated with calving ease. As birth weight goes up, calving ease tends to decrease. The genetic correlation between calving ease direct and birth weight in American Angus (https://www.angus.org/Nce/Heritabilities.aspx) is -0.65, which means that birth weight explains 42% of the genetic variation in calving ease (-0.65*-0.65=0.42), with the remainder being determined by other factors unique to calving ease. Calving ease direct and calving ease maternal also share a small, negative genetic correlation. These two calving ease traits have an antagonistic relationship, which means that direct calving ease should be used when selecting bulls to use on heifers, but that maternal calving ease should be the preferred selection metric for choosing sires of replacement females.
|Directionality||Trait 1||Trait 2||Magnitude|
|Positive||Milk production||Total Energy Intake||Moderate|
|Positive||Fat Thickness||Yield Grade||High|
|Positive||Carcass Weight||Birth Weight||
|Positive||Yearling Weight||Mature Size||Moderate to High|
|Positive||Calving Interval||Milk Yield||High|
|Negative||Birth Weight||Direct Calving Ease||High|
|Negative||Direct Calving Ease||Maternal Calving Ease||Low to Moderate|
|Negative||Average Daily Gain||Feed to Gain Ratio||Moderate to High|
|Uncorrelated||Age at Puberty||Retail Product (%)||Negligible|
|Uncorrelated||Weaning Weight||Bone Percentage||Negligible|
As one might expect, carcass attributes that tend to aggregate a variety of phenotypes, such as quality grade and yield grade, are correlated to their component traits. For example, marbling and quality grade have a positive favorable correlation, and yield grade and ribeye area have a negative favorable correlation. Carcass traits also share favorable genetic correlations with ultrasound measurements of carcass merit, which means that ultrasound measures on yearling seedstock animals can be utilized as indicator traits for genetic evaluation of carcass merit.
Genetic correlations also exist for traits within the cowherd. For example, milk production and maintenance energy have an antagonistic relationship. Antagonistic relationships also exist between growth traits and mature size. As increased growth is selected for in calves within a herd that keeps replacement females, mature size can inadvertently increase, which increases maintenance energy in the cowherd. These examples all show that selection rarely happens in a vacuum, so it is important to be aware of all of the traits that are relevant in a particular herd and jointly select for merit in all the economically relevant traits in the breeding objective to overcome the effects of genetic antagonisms.
Causes of Genetic Correlations
Genetic correlations can have several different causes. The first of those is pleiotropy, the case where one gene impacts multiple traits. Sometimes this gene acts favorably on two or more traits, and sometimes it creates an antagonism. An example of this is the DGAT1 gene. One of the alleles (form of a gene) for DGAT1 increases milk fat, but has a negative impact on milk yield. In a population highly selected for milk yield, we would expect this allele to be at low frequency. In a population highly select for milk fat, we would expect the opposite to be true. These types of relationships that contribute to genetic correlations cannot be “broken” or separated because it’s a single gene impacting two or more traits. However, genetic correlations can also be caused by genes that are close together on the same chromosome. Because they are close together on the same chromosome, they tend to be inherited together. As a favorable allele for one trait is selected for, the alleles next to it also get selected for and tend to be inherited together. These types of genetic correlations (or genetic antagonisms) can be “broken” or separated over time by selecting for genetic merit in the two traits simultaneously. Another option is to utilize selection indexes, which can break up genetic antagonisms, or unfavorable genetic correlations, over time due to their emphasis on appropriate weighting for all traits (including those with antagonistic relationships) in one selection tool.