meta Dairy Genomic Indexes Stand The Test of Time :: The Bullvine - The Dairy Information You Want To Know When You Need It

Dairy Genomic Indexes Stand The Test of Time

A major new study of Holstein dairy sires has confirmed that genomic indexes are a remarkably reliable predictor of actual daughter performance. This provides reassurance to the thousands of UK dairy producers who select service sires on the strength of an index based on their DNA.

Our study included every genomic sire marketed in the UK in August 2014 and compared its genomic index at that time with its daughter-proven index in August 2018.  

Some 7,745 bulls were included in the analysis which showed there was little difference between their average genomic index in 2014 and their proven index four years later.

Marco Winters, head of animal genetics for AHDB Dairy says: “This study should give complete reassurance that UK genomic indexes are amongst the most stable and reliable in the world. Provided they chose their service sires wisely, the many producers who put their faith in genomic indexes should have made substantial genetic gains in their herds.”

We introduced genomic indexes for bulls marketed in the UK in 2012. At that time, the technology to calculate the indexes was still in its early stages, and producers were made aware they were more likely to change over time than indexes for proven sires which already had many daughters milking.

However, the attraction of using these young sires was their generally superior genetics which would help producers breed better replacement dairy heifers. Thousands of producers took this opportunity and some now breed 100 per cent of their replacements from genomic sires. Across the UK, around 70 per cent of dairy replacements are now bred from these young bulls.

Full results from the study group are shown in Table 1. The correlation figure, also in the table, gives an indication of how close all of the sires are to that average change figure. Correlation figures for all traits were generally considered high, indicating there was a narrow spread of bulls around the average. In other words, as well as the average index in 2018 being close to, or the same as, the average in 2014, there was not a wide spread of bulls around that average. This means that most individual bulls’ indexes were close to their former prediction. The graph depicts this information in figures and indicates that between 2014 and 2018, 34.2 per cent of bulls’ indexes (in this case for Lifespan but equally applicable for other indexes) did not change at all and most others have changed very little.

A second analysis of a smaller sub-set of bulls which now have far more daughters showed exactly the same pattern. This gives further reassurance that as large numbers of daughters finally contribute to a bull’s index later in life, their indexes are still unlikely to see much change on average from their initial genomic predictions.

Marco continued: “We were confident in the system when we launched genomic indexes in the UK in 2012, and this analysis confirms the reliability of the techniques we used.”

Remarking that the UK teamed up with the USA, Canada and Italy to provide one of the largest reference populations in the word, he said this helped ensure the reliability of the UK indexes.

Table 1: Aug ’18 PTAs for daughter-proven sires compared with their genomic indexes in Aug ’14

Predicted Transmitting Ability (PTA) Aug ’18 (average) Change since 2014 Correlation between Aug ’14 and Aug ’18
Reliability (milk, fat and protein) % 83 15  
Milk kg 452 30 0.85
Fat kg 15.0 -0.6 0.85
Protein kg 13.7 0.1 0.86
Fat % -0.03 -0.02 0.90
Protein % -0.01 -0.01 0.87
Reliability (SCC) 83 14  
SCC -11 0 0.83
Reliability (Lifespan) 69 9  
Lifespan 0.25 -0.01 0.80
Reliability (Fertility Index) 72 9  
Fertility Index 3.2 0.3 0.85

 7,745 sires contributed to the data in this table

 Graph 1 Distribution Of Percentage Change Of Genomic Indexes


Graph 1: Distribution of percentage change between young sires’ genomic indexes for Lifespan PTA in August 2014 and the same sires’ daughter-proven indexes in August 2018


Genomic reminder

What’s a genotype?

The genotype is the genetic make-up (or gene-set) of the animal. It is measured from a tissue sample containing the animal’s DNA (its genetic blueprint) and expressed as a raw code which goes on to be interpreted by AHDB Dairy, through its contract partners, Egenes, and converted to a genomic index.

What’s a genomic index?

A genomic index is a genetic index which is calculated from the gene-set or genotype of a young animal, rather than from its own, its daughters’ or its parents’ and ancestors’ actual performance. In reality, the genomic component of an animal’s index is used alongside other genetic information from these alternative sources, in a smooth transition towards an index which will ultimately be based on actual or daughter performance.

What’s a reference population?

A reference population (also called a training population or a predictor population) comprises several thousand animals whose phenotype (actual performance/appearance) and genotype (or gene-set) are compared. This allows relationships to be established and from these, genomic Predicted Transmitting Abilities (PTAs) are calculated.

What do the correlations mean?

Correlation figures indicate how widely spread the population is around the average (in this case, the mean). To demonstrate with the extremes, a correlation of zero would indicate that the spread around the average was huge – in other words, while the average change in index for all sires may have been small, large numbers of bulls would not have been close to their former prediction.

Conversely, a correlation of one would mean that every bull’s daughter index was identical to its former genomic index. Obviously that would be an impossibility, and wouldn’t even happen between index runs for a group of daughter-proven sires!

What’s the recommended usage of genomic sires?

Genomic indexes remain marginally less reliable than daughter-proven indexes which means their likelihood of change is greater. However, this study has indicated the indexes used over the past four years have been extremely accurate. Ultimately, farmers have to take their breeding decisions based on their attitude to risk. Usage should continue to be based on the existing advice, which is to limit the use of any single young sire to 12.5% of the herd, which – if all inseminations are to genomic sires – means using a minimum of eight young sires each year. Current young sire usage runs at about 70% of dairy inseminations in the UK.


Source: AHDB Dairy

Send this to a friend