@2017 by Performance Genetics LLC

How inbreeding affects racing performance in Thoroughbreds

April 24, 2018

Authored by -

 

An interesting paper was released as an Open Access (which means anyone can read it without paying) on how inbreeding affects racing performance in Thoroughbred racehorses in Australia.

 

The study looked at the performance of 135,572 racehorses in Australia that had one or more race starts between 2000 and 2011. For all purposes these were horses probably born between 1995 and 2005 which is important to note as it was during this period that Australian breeding saw the largest influx of overseas stallions into their breeding population - in addition to names like Danehill, Anabaa, Red Ransom, Dehere and Royal Academy, think of names like Fuji Kiseki, Sandpit, Bubble Gum Fellow, Piccolo, Gilded Time, Shalford and Brocco - all stallions with more diverse pedigrees that served commercial mares at the same time. 

 

There were a couple of interesting finds in the study. The first of which was the relationship between Wrights Inbreeding Coefficient (F) and Baumung, et al's Ancestral History Coefficient (Ahc) and some common metrics of performance. 

 

 

The measure of F, or the inbreeding coefficient of the horse, had a negative relationship with all of the performance metrics - that is, the more inbred a horse was on paper, the less ability it seemed to have. This is not a unique thought or finding as it applies in other breeds as well and Steve Harrison over at Thoroughbred Genetics has for some time stating  that there is an optimum amount of gene duplication (inbreeding) that occurs in the best performers. 

 

Conversely, the relationship between the Ancestral History Coefficient and these performance metrics is a positive one. The authors of the paper note that the Ahc is a reflection of the selection of favorable performance traits over time, that is, horses with a higher Ahc are more likely to contain larger proportions of alleles in their genomes that have been positively selected over many generations. 

 

Another interesting find in the paper was the specifics behind the uneven ancestral genetic contribution in the contemporary population. The thoroughbred is built upon assortative matings, that is, we believe as humans that we are the best arbiters of how to improve the breed (where if nature was left to its own devices it may select a sire-dam mating differently), so over a period of time we have favored ancestors that provide what we deem to be positive genetic material which by default has lead to the extinction of ancestors determined to have less than positive genetic material (just think for a minute how a bad stallion like Catrail who was serving mares in the 1990's has quickly disappeared from the commercial breeding population and is rarely seen even as a broodmare sire).

 

 

The authors noted 10 ancestors who accounted for over 82% of the Ancestral History Coefficient which would indicate that these 10 ancestors had a highly impactful influence on the breed. 

 

More importantly, in using a linear mixed model to look at the relationship between partial inbreeding coefficients and racing performance they found that there was a uneven distribution of genetic load between ancestors, that is, inbreeding to a particular ancestor results in a reduction of racing performance.

 

 

The authors found that inbreeding to four ancestors had significant effects on racing performance. Individuals who were more inbred to Herod had greater cumulative earnings, earnings per start and career length. Conversely, inbreeding to Eclipse, Stockwell and Touchstone (in particular) had negative effects on the racing performance of their descendants. 

 

Interesting stuff.

 

There is of course a logical step to make from this to further test the findings of this paper and see if it has a real world application. If F and Ahc are associated with performance, as is the presence of Herod, Eclipse, Stockwell and Touchstone in pedigrees, it should be fairly straight forward to set up a little study to test their combined relationship with performance. A simple logistic regression to start with by setting up the following for say 1,000 horses:

 

  • Performance (1 = elite runner; 0 = non-elite; this would be the target variable)

  • F

  • Ahc

  • Count of appearances of Herod in pedigree

  • Count of appearances of Eclipse in pedigree

  • Count of appearances of Stockwell in pedigree

  • Count of appearances of Touchstone in pedigree

  • Pedigree Score (give a +1 for the number of appearances of Herod and a -1 for appearances of Eclipse, Stockwell & Touchstone)

  • Weight of appearances of Herod in pedigree (you'd work out the percentage of inheritance based on the generation the name appeared in and add them up)

  • Weight of appearances of Eclipse in pedigree

  • Weight of appearances of Stockwell in pedigree

  • Weight of appearances of Touchstone in pedigree

 

If you then looked at the stallions that served the most mares in Australia from 1995-2005 and took all their graded/group stakes winners sired during that period, and then selected the same number of horses by those sires that had 5 or more starts and couldn't beat a fat man down a hill, taking out full relations, you would then have a decent population to look at the effect of these variables and to see if anything predictive could be made.

 

Something to tackle this summer..... 

 

 

 

 

 

 

 

 

 

 

 

 

 

Please reload

Featured Posts

Mitochondrial DNA haplotypes and mating outcomes

September 3, 2018

1/8
Please reload

Recent Posts
Please reload

Archive