The ins and outs of indexes

Acronyms like TPI, NM and MFP roll off the tongues of AI representatives like the sportscasters who read stats from your favorite baseball team. But, if you don't have a good understanding of these indexes, you may find yourself being just as lost as child who watches his first baseball game and hears about a pitcher's ERA.

With more than 20 statistics calculated for every bull by the USDA and Holstein Association in the sire summaries, it is easy to get lost and not recall how each value is derived. However, understanding the basics of indexes and whether they are right for your breeding program is critical to reaching your goals.

To help you stay on top, we have analyzed four indexes. Two of the indexes are familiar ones - the Holstein Association's Type Production Index, or TPI, and the Net Merit Index, or NM. The other two were just introduced last month - the Cheese Merit Index (CM) and the Fluid Merit Index (FM). The new formulas, developed by geneticists at USDA's Animal Improvement Programs Laboratory (AIPL), are "sister" formulas to NM.

Trait definitions and formulas that make up the indexes are as follows:

Type Production Index
Developed more than 20 years ago, the TPI index is one of the longest running indexes. The current formula for TPI includes the following ratio: 3 PTA protein: 1 PTA fat: 1 PTA type: 0.65 udder composition: 0.35 feet and legs composite.
Components of the TPI index:

PTA: PTA stands for predicted transmitting ability. This is the best estimate of the sire's superiority (or inferiority) in transmitting production or type characteristic to daughters. The genetic base is established every five years from production and type data recorded by DHI and breed associations.

PTA protein: Measures the PTA for producing protein.

PTA milkfat: Measures the PTA for producing milkfat.

PTA type: PTA type is a composite of all type traits evaluated by the Holstein Association's classifiers, including udder traits, feet and legs characteristics as well as the structure of the cow - frame, body capacity and dairy character.

Udder composite: The udder composite is a compilation of udder traits:
(udder depth) (0.3) + (teat placement) (0.16) + (udder height) (0.16) + (udder width)(0.12) + (udder cleft)(0.1) = udder composite.

Feet and Legs composite: The feet and legs composite looks at traits using the following formula:
0.5 x [(foot angle)(0.48) + (rear legs/rear view)(.37) - (rear legs/side view)(0.15)] + 0.5 (feet and leg score) = feet and legs composite.

How the TPI index is calculated:
[(3)(PTA protein/19) + (PTA milkfat/22.5) +(PTA type/0.7) + (0.65)(udder composite/0.8) + (0.35)(feet and legs composite/0.85)]50 + 576 = TPI

Observations and differences: The TPI differs from the other indexes listed below because it directly includes type data. According to Tom Lawlor, director of research and development for the Holstein Association, type data is used in the index to predict an animal's longevity. Conversely, the NM, FM and CM use PTA productive life (PL) - an estimate in months of how long a sire's daughters will remain in the herd after freshening - to measure longevity.

"There are two ways to measure longevity," Lawlor says. You can look at the traits, such as udders, feet and legs, that influence whether a cow stays in the herd or you can measure how long she actually stays.

Analyzing type traits provides you with an advantage because you can actually see the traits improve or decline, says Lawlor. And, most type traits have a higher heritability than PTA productive life. For example, udder traits range in heritability from 23 percent to 29 percent, while productive life has a heritability of 8.5 percent.

"When TPI was started, we didn't have PTA productive life to use for longevity," Lawlor says.

TPI also differs from the three other indexes in that it does not include PTA milk. Lawlor says PTA milk was replaced with PTA protein over the years as protein took on more importance in milk pricing. However, he notes the two traits - PTA protein and PTA milk - have a correlation of 0.8. That means sires with high values for protein will usually have high values for milk pounds.

Net Merit Index
Developed in 1994, the net merit index (NM) incorporated longevity and somatic cell information into an existing production-related index known as milkfat protein dollars (MFP$). It was the first attempt by USDA-AIPL researchers to incorporate a measure for longevity in the sire summaries.

The net merit index weights MFP$, PTA productive life, PTA somatic cell score (divided by their standard deviations) in a ratio of 10:4:1.
Components of the NM index:
Dollar values used in indexes: The current values are $12.30 per hundredweight for milk, 3.5 percent protein and 3.2 percent milkfat.

MFP$: Milkfat protein dollars is formulated as: $0.031 (PTA milk) + $0.80 (PTA milkfat) + $2 (PTA protein).

MFP$ feed cost: USDA-AIPL geneticists estimate that feed cost represents nearly 30 percent of the MFP. Therefore, the value is calculated: (MFP$)(0.3).

PTA PL: Productive life measures the months a cow produces milk between the time she freshens until she reaches the age of seven. A maximum of 10 months are calculated per lactation. A cow must reach the age of three for its productive life data to be included in the calculation. Until that time, geneticists use an estimate based on the productive life of the cow's sire and dam.

PTA SCS: Geneticists derive somatic cell scores from the linear somatic cell counts (SCC) that you see as a part of your DHI test. Although somatic cell counts do not directly indicate a cow's treatment history for mastitis, the SCC has a correlation with mastitis treatment of approximately 0.6.

Breed average for SCS: Geneticists subtract each individual bull's SCS value from his respective breed average. The breed averages for SCS are:

  • Holstein 3.2
  • Jersey 3.3
  • Brown Swiss 3.22
  • Ayrshire 3.15
  • Guernsey 3.35
  • Milking Shorthorn 2.88

How the NM index is calculated:
NM = MFP$ - MFP$ feed costs +$11.30(PTA PL) -$28.22(PTA SCS - breed average for SCS)

Observations and differences:
Dairy producers often have difficulty grasping the meaning of the economic values in the indexes, says Jay Mattison, executive vice-president of ReQuest, Ltd, an information firm in Madison, Wis., that publishes sire information on the Internet.

Obviously, some predicting must be done to develop price and cost data.. Duane Norman, geneticist at USDA-AIPL, acknowledges that when working on indexes, geneticists often project milk prices further into the future than most dairy economists.

Why aren't the prices changed more frequently? Ron Pearson, dairy geneticist at Virginia Tech, describes the paradox. "While we want indexes that reflect the market prices, we don't want to change the formula so much that bulls are rising and falling on the list throughout the year."

These observations remind producers that the absolute value of the index may not be as important as how the bulls rank relative to one another. "Even if you don't agree with the dollar values in the calculation, the formula treats every bull the same," says Mattison.

Another important note for producers using NM is to keep the reliability of PTA PL and PTA SCS in proper perspective. These particular traits do not have the same degree of heritability and reliability that milk production traits have.

PTA PL has a heritability of 8.5 percent, while milk has a heritability of 30 percent. Reliability for PL is around 50 percent for most bulls and 70 percent or better for milk production. PTA SCS has a heritability of 10 percent and a reliability between 55 and 60 percent.

Bottom line: Genetic improvement for PL and SCS takes time and big changes in longevity or somatic cell counts should not be expected.

Fluid Merit Index
With the February 1999 sire summaries, the USDA-AIPL released two new indexes. The first, Fluid Merit Index (FM), replaces the milkfat dollar (MF$) values and includes longevity and somatic cell information. As the name implies, the index was designed for "a producer who expects to receive no income from protein" when selling his milk, says Paul VanRaden, dairy geneticist at USDA-AIPL.

The index is calculated at: 10 MF$: 4 PTA PL: 1 PTA SCS.

Components of the FM index:
MF$: Milkfat dollars is calculated in the following manner: $0.095(PTA Milk) + $0.80(PTA fat).

All other components of the index are the same as in NM.
How the FM index is calculated:
FM = MF$- MFP$ feed costs + $11.30(PTA PL) -$28.22(PTA SCS - breed average for SCS)

Observations and differences:
The FM differs from the three other indexes because it contains no economic incentive for protein. In fluid-based markets, protein does not enter the equation because "fluid processors do not standardize milk for protein," says Joe Hillers, professor of animal science at Washington State University. For example, if you purchase a gallon of milk at the store with 1 percent fat, you know the milkfat value within about 0.5 percent, says Hillers. The protein content of the same milk could vary from 2.8 percent to 3.2 percent. "You have no idea what the protein is."

However, a potential concern is that producers selecting solely on milk will end up with too low of a protein content. While fluid producers don't receive any incentive for protein, says Pearson, "you don't want to be selling a product that's basically water, either." A low protein level could eventually lower the solids-non-fat level of the milk, which could penalize the producer economically.

Right now, that's not much of a concern due to the high genetic correlation - 0.8 - between milk and protein. But, if the value of protein in high-ranking FM bulls declines, then fluid producers should consider some form of protein selection.

Additionally, the index places an indirect negative weight on protein by including the feed cost for protein even though the producer receives no premium for protein. Research has shown that it takes more feed to produce a pound of protein than a pound of fat or milk, says VanRaden. So, by subtracting the feed cost associated with protien from the MF$ value, the index's total value is less than NM and CM.

Cheese Merit Index
The Cheese Merit Index (CM) is also a new introduction of the USDA-AIPL for the February 1999 sire summaries. The CM replaces the cheese yield dollar (CY$) value and provides longevity and somatic cell score data for producers whose milk is made into cheese.

The ratio for the CM is similar to FM and NM: 10 parts CY$: 4 parts PTA PL: 1 PTA PL.
Components of the FM index:
Cheese Yield Dollars (CY$): The equation is: -$0.002218 (PTA milk) + $0.80 (PTA milkfat) + $3.1876 (PTA protein) = CM.

All other components are the same as NM.
How the CM index is calculated:
CM = CY$- MFP$ feed costs + $11.30(PTA PL) -$28.22(PTA SCS - breed average for SCS)

Observations and differences:
CM differs from FM and NM in that it contains a negative value for milk. The FM uses a value of $0.095, while the NM uses a value of $0.031.

In other words, CM targets producers who have their milk priced under the multiple-component pricing system where the price for pounds of milk is lower than the price for milkfat or protein. About 85 percent of the U.S. milk will be priced under multiple-component pricing if the federal order reform process - expected to be complete by October - stays on track. The southeastern U.S. and Arizona will not be under multiple-component pricing.