Microbiome features as future traits
Information could improve the accuracy of heifer pregnancy EPD at a relatively early age.
By Andrew D. Lakamp and Matthew L. Spangler
The suffix “-ome” is Greek meaning all or the totality of something. The “ome” that most producers are likely aware of is the genome, or all the genetic material an organism possesses. However, other useful “omes” exist. A microbiome is a collection of all the microbes that live in a specific place, like the rumen. Because a microbiome is made up of many components, researchers use the term feature to describe specific aspects of it.
The definition of a feature depends on the context of the study and can range from a single microbial gene to a particular microbial species to broader measures like overall microbial diversity. For simplicity, here we use “feature” to mean a specific DNA sequence - essentially a distinct set of biological instructions. It is these features of the microbiome which could serve as phenotypes that producers can actually select on. However, before such action can be taken there are numerous questions that need to be answered.
For a trait to have any value for genetic selection it must meet two criteria: 1) it must be heritable, and 2) it must either be an economically important trait or serve as an indicator for one. Ideally, indicator traits would be easier/cheaper to measure and would be collected earlier in life.
Microbiome features themselves are not economically important as there is no source of revenue or cost directly associated with them. However, several studies have shown the measure of how common a specific feature is relative to all features, i.e., the relative abundance of a feature, is both heritable and genetically correlated with economically important traits.
For example, work done at the University of Nebraska-Lincoln has identified multiple rumen microbiome features which are heritable and genetically correlated with average daily dry matter intake (ADDMI) and average daily gain (ADG). Even more interesting, one heritable feature was positively correlated with ADG and negatively correlated with ADDMI. This means selecting for an increased proportion of that feature would indirectly select for improved feed efficiency by increasing ADG while also decreasing ADDMI.
Most routinely collected phenotypes are only measured once (e.g., weaning weight) or a few times during animal’s life (e.g., pregnancy status). In comparison, the microbiome provides thousands to millions of features to potentially use as phenotypes for each sample taken and multiple samples can be collected over time.
Current research suggests the vast majority of these features are either not common enough to focus on, not heritable, or not correlated with an economically important trait. However, the volume of data still leaves plenty of candidates for selection which do meet all those criteria.
There are examples in the scientific literature that show proof of concept of using selected microbial features to improve the response to selection. For example, geneticists from the Scottish Rural College (SRUC) in Edinburgh used 30 microbiome features with genetic correlations to methane and found selecting on just the microbiome features resulted in 22 – 34% greater reductions in methane than if they had used methane measurements alone. It is worth noting, however, that by using both methane measurements and the microbiome features, methane decreased by another 21 – 22% over using just the microbiome features.
One of the most useful ways to use microbiome features for selection might be to increase the expected progeny difference (EPD) accuracy for lowly heritable traits like fertility, or for sparsely collected traits like methane emissions and feed intake. Researchers from the University of Arkansas were able to identify a group of 15 vaginal microbiome features taken from heifers before breeding that were predictive of heifer pregnancy. They also identified three features in particular which were more abundant in open heifers compared to bred heifers.
While no work has been done in beef cattle to calculate the heritability of vaginal microbiome features, research conducted at Iowa State University showed the relative abundances of some vaginal microbiome features were heritable in gilts. A thorough investigation has yet to be conducted, however it may be possible that features of the vaginal microbiome taken pre-breeding are both heritable and genetically correlated with fertility traits such as heifer pregnancy. If that were the case, then selection on the microbiome features would not necessarily be required for them to be valuable. Using microbiome features as additional, correlated pieces of information could improve the accuracy of the heifer pregnancy EPD at a relatively early age.
As the cost of genetic sequencing decreases, more and more microbiome data is becoming available in research settings. While interesting from a pure science perspective, these can have applications in cattle breeding by potentially becoming new phenotypes. Microbiome features, if heritable and genetically correlated with economically important traits, can provide new selection targets as indicator traits to help improve EPD accuracy for traits that are hard or expensive to measure or occur late in life.
The use of data from the microbiome for improving genetic selection in beef cattle is still in its infancy and there are many unanswered questions. However, this is an exciting new arena and there is no doubt that cattle producers will continue to hear more about the microbiome and its effect on traits that matter to them.
Lakamp was a former post-doctoral research associate and Spangler is the Ronnie D. Green Professor of Animal Science and a beef genetics extension specialist, both in the Department of Animal Science, at University of Nebraska-Lincoln.