GENETIC SELECTION FOR MILK QUALITY WHERE ARE WE? David Erf Dairy Technical Services Geneticist Zoetis

GENETIC SELECTION FOR MILK QUALITY WHERE ARE WE? David Erf Dairy Technical Services Geneticist Zoetis

OVERVIEW» The history of genetic evaluations» The importance of direct selection for a trait» Selection for improving mastitis today» How the information is working in the field

HEADLINES: Holstein is invisible cow': Trouble-free animal nears lifetime milk record in just 10 lactations Betzoldvale Scott Mar. 11-year-old Holstein has produced 348,450 pounds of milk in her lifetime (as of May 2011) Became pregnant on her first service insemination every year www.thecountrytoday.com/front_page/article_ec780704-86d1-11e0-8571- 001cc4c002e0.html Accessed January 2, 2016

RECOVERING HEIFER RAISING COSTS TAKES TIME Adapted from Ferguson and Galligan, WCDS, 1995.

Milk/Cow/Day using 305M % of Total LONGEVITY-DRIVEN PROFIT 115 110 105 100 95 90 85 80 75 70 1 2 3 4 5 6 7 8 Parity Group Data on file, Dec. 2015 Herd example, Zoetis, Inc. Performance and Percent of Herd by Parity Cows >100DIM Milk/Day % of Herd 45% 40% 35% 30% 25% 20% 15% 10% 5% 0%

IMPACT OF MORBIDITY IS SIGNIFICANT IN DAIRY COWS Incidence per Lactation Range Cost ($) per Case Culling Risk 1 (%) Displaced Abomasum 3-5% 1,2,3,4,13 $494 4 26.9 Ketosis 5-14% 1,3,4,13 $117-289 4,5 32.5 Lameness 10-48% 2,4,6,13 $177-469 4,7 16 2 Mastitis 12-40% 1,2,3,4,8,13 $155-224 4,8,9 32.7 Metritis 2-37% 1,3,10,11,13 $300-358 10,11 17.1 Retained Placenta 5-15% 1,2,3,4,11,12 $206-315 4,12 31.7 1 Grohn, Y. et al. 1998. Effect of Diseases on the Culling of Holstein Dairy Cows in New York State. J. of Dairy Sci, 81(4):966-978. 2 USDA. 2008. Dairy 2007, Part II: Changes in the U.S. Dairy Cattle Industry, 1991 2007 USDA-APHIS-VS, CEAH. Fort Collins, CO #N481.0311. 3 Bar, D., et al. 2007. Effect of repeated episodes of generic clinical mastitis on milk yield in dairy cows. Journal of dairy science 90(10):4643-4653. 4 Guard, C. 2009. The costs of common diseases of dairy cattle. Central Veterinary Conference Proceedings. Kansas City, MO.. 5 McArt, J.A. et al, 2015. Hyperketonemia in early lactation dairy cattle: a deterministic estimate of component and total cost per case. J. of Dairy Sci. 98(3):2043-2054. e Bicalho, R.C. Lameness in Dairy Cattle: A debilitating disease or a Disease of Debilitated Cattle? Western Dairy Management Conference, 2011. Pg 73-83. 7 Cha, E. et al. 2010. The cost of different types of lameness in dairy cows calculated by dynamic programming. Preventive veterinary medicine 97(1):1-8. 8 Cha, E, et al. 2014. Optimal insemination and replacement decisions to minimize the cost of pathogen-specific clinical mastitis in dairy cows. Journal of dairy science 97(4):2101-2117. 9 Cha, E, et al. 2011. The cost and management of different types of clinical mastitis in dairy cows estimated by dynamic programming. Journal of dairy science 94(9):4476-4487. 10 Overton, M. and J. Fetrow. 2008. Economics of postpartum uterine health. Proc Dairy Cattle Reproduction Council:39-44. 11 "The Value of Uterine Health: the diseases, the causes, and the financial implications." Dairy Cattle Reproduction Council article. 12 Guard, C., 1999. Retained Placenta: Causes and Treatments. Advances in Dairy Technology (1999) Vol. 11, page 81. 13 Zwald, N.R., K.A. Weigel, Y.M. Chang, R.D. Welper and J.S. Clay. 2004. Genetic selection of Health Traits using Producer-Recorded Data. I. Incidence Rates, Heritability Estimates and Sire Breeding Values. J. of Dairy Sci., 87:4287-4294.

THE HISTORY OF NEW TRAITS IN GENETIC EVALUATIONS» Prior to 1994 Milk, Fat, Protein and Type» 1994 Net Merit (NM$), SCS, Productive Life (PL)» 2003 Daughter Pregnancy Rate (DPR)» 2008 Stillbirth Daughter and Sire» 2011 Heifer Conception Rate (HCR), Cow Conception Rate (CCR)» 2016 - Livability

WELLNESS IS NOW A PROFITABLE CHOICE

CLARIFIDE PLUS FOR HOLSTEINS CDCB Official Evaluation Parentage Production Reproduction Health Type Wellness Traits Mastitis Lameness Metritis Retained Placenta Displaced Abomasum Ketosis Genetic Conditions Polled (no fee) Milk Components Infertility Haplotypes Other genetic conditions* DWP$ TM Animal Ranking * CVM, Brachyspina and Beta Casein A2 available with add-on fee.

CREATING WELLNESS TRAIT GENOMIC PREDICTIONS

PHENOTYPIC MEANS, STANDARD DEVIATIONS, AND HERITABILITIES USED IN THE ANALYSES Trait Unit Mean SD h 2 Mastitis Lameness Metritis Retained placenta Displaced abomasum Ketosis Case per animal/lactation Case per animal/lactation Case per animal/lactation Case per animal/lactation Case per animal/lactation Case per animal/lactation 0.23 0.42 0.069 0.20 0.40 0.063 0.10 0.30 0.059 0.05 0.22 0.073 0.02 0.15 0.081 0.05 0.22 0.059 Each trait was defined as a Holstein female recorded with the presence or absence of a disease/disorder in a given lactation Every cow was assigned with 0 (no incidence) or 1 (recorded one or more incidences) per lactation Source: Data on file, Zoetis internal data, August 2015, Zoetis Inc.

MILLIONS OF RECORDS CONTRIBUTE TO AVERAGE RELIABILITIES OF 49 OR HIGHER Trait Reliabilities (%) of young genotyped and pedigreed females GPTA Reliability Traditional Parent Average Reliability MAST 51 19 LAME 50 18 METR 49 17 RETP 50 17 DA 49 16 KET 50 16 29,901 heifer observations (<2 years old) Reliabilities are similar or better than Reliability of some core CDCB traits (i.e. HCR & DSB) Source: Data on file, Zoetis internal data, August 2015, Zoetis Inc.

CLARIFIDE PLUS GENOMIC PREDICTIONS» Wellness traits are expressed as a standardized transmitting ability (STA) Centered at 100 with a standard deviation of 5 Higher numbers represent animals with lower expected risk of disease relative to herdmates with lower STA values Generally range from 75 to 120 for all traits +/- 2 SD

SOMATIC CELL SCORE FOR HOLSTEIN S HAS BEEN IMPROVING OVER TIME Genetic Progress for SCS (heritability = 0.12) Data on file, Enlight herd data, Accessed Dec 2015, Zoetis Inc.

YET DAIRY COW MORBIDITY HAS NOT BEEN IMPROVING FOR MASTITIS OR OTHER TRAITS 70 Disease Trends 60 50 40 30 20 10 Displaced Abomasum Other Reproductive Problems (e.g. dystocia, metritis) Infertility Problems (not pregnant 150 days after calving) Retained Placenta Lameness Clinical Mastitis 0 1996 2001 2007 USDA. 2008. Dairy 2007, Part II: Changes in the U.S. Dairy Cattle Industry, 1991 2007 USDA-APHIS-VS, CEAH. Fort Collins, CO #N481.0311.

GPTA Protein [lbs] SOME TRAITS HAVE HIGH CORRELATION BETWEEN EACH OTHER, LIKE MILK AND PROTEIN Yet selecting for Protein, to get more protein, is best 70 Correlation in Breed = 0.83 a Observed Correlation=0.81 60 50 40 30 20 10 0-500 0 500 1000 1500 2000 GPTA Milk [lbs] a http://aipl.arsusda.gov/reference/nmcalc-2014.htm, Accessed Dec. 15, 2015 Data on file, Dec. 2014 Data package n = 30, Zoetis Inc.

SOME TRAITS HAVE MEDIUM OR LOW CORRELATION, LIKE MILK AND FAT GPTA Fat [lbs] Here direct selection for Fat is highly necessary if you want to improve fat Correlation in Breed = 0.43 a Observed Correlation=0.20 100 90 80 70 60 50 40 30 20 10 0-500 0 500 1000 1500 2000 GPTA Milk [lbs] a http://aipl.arsusda.gov/reference/nmcalc-2014.htm, Accessed Dec. 15, 2015 Data on file, Dec. 2014 Data package n = 30, Zoetis Inc.

BOTTOM LINE, DIRECT SELECTION IS BEST!» To make the most genetic and phenotypic progress in a trait, use direct selection, not indirect» Examples: Goal Indirect trait Direct trait Estimated correlation Improve reproduction PL Daughter Pregnancy Rate (DPR) 0.64 Reduce mastitis cases Reduce lameness SCS New genomic Mastitis trait -0.45 PL New genomic Lameness trait 0.28 FLC New genomic Lameness trait 0.00 Reduce metritis PL New genomic Metritis trait 0.32 Source: Data on file, Zoetis internal data, August 2015, Zoetis Inc.

Average Disease Incidence DIRECT SELECTION IS BEST! Genomic Mastitis vs GPTA for SCS Data includes Ref. Pop. n= 76K Not useful for validation purposes since phenotypic records are contributing to the Mastitis Prediction Good example we can expect from the validation data of how direct selection is best Genomic Prediction Quintiles Source: Data on file, Zoetis internal data, August 2015, Zoetis Inc.

STUDY TIMELINE August 2016 Health Records collection Year 1 Year 2 Year 3 September 2015 Enrollment & DNA Sample collection prior to calving Oct 2016 Oct 2017 Oct 2018

DISTRIBUTION OF ENROLLED HERDS

LOCATION AND SAMPLING OF ANIMALS Herd State Number of Sampled Animals 1 MI 318 2 MN 308 3 ID 344 4 CA 350 5 CA 350 6 WI 292 7 NY 253 8 GA 305 9 CA 334 10 WI 363 11 ID 350 Animals Sampled 3567 Animals eliminated* -692 Animals included in the study 2875 (66% Nulliparous, 33% Primiparous) * Animals eliminated due to calving date outside desired window, breed conflict, sample failures Source: Data on file, Zoetis internal data, October 2016, Zoetis Inc.

DEFINING AFFECTED AND HEALTHY RECORDS First and second lactation records were assigned to one of three possible phenotypes by trait: 1. Healthy (0): no documented incidence of the health event before 306 DIM 2. Diseased (1): a documented incidence of the health event before 306 DIM 3. Excluded (.): missing lactation record and/or animals that were healthy and sold/died before the 90 th percentile DIM by trait RP 4 DIM KET DIM 16 MAST DIM 250 Calving DIM 0 305 Day Lactation METR DIM 9 DA DIM 66 LAME DIM 265

Marginal Means LS-Meas STA ASSOCIATION BETWEEN MAST STA GROUPS AND MASTITIS EVENTS 20.0% Prevalence: P<0.0001; STA: P<0.0001 110 15.0% 15.9% A 101.1 Y 105.2 Z 105 11.2% B 11.1% B 100 10.0% 97.7 X 8.5% C 95 5.0% 92.3 W 90 0.0% Worst 25% 26-50% 51-76% Best 25% Genetic Groups 85 Source: Data on file, Zoetis internal data, October 2016, Zoetis Inc.

Marginal Means LS-Meas STA ASSOCIATION BETWEEN MAST STA GROUPS AND MASTITIS EVENTS 20.0% 15.0% 10.0% 5.0% 0.0% 15.9% A 92.3 W Prevalence: P<0.0001; STA: P<0.0001 11.2% B 11.1% B 97.7 X 101.1 Y Genetic Group Est. Cost per Cow 5 Bottom 25% $35.74 26-50% $25.07 51-75% $24.71 Top 25% $18.81 5 Cha et al., 2011. Journal of Dairy Science 94(9):4476-4487. 105.2 Z 8.5% C Worst 25% 26-50% 51-76% Best 25% Genetic Groups 110 105 100 95 90 85 Source: Data on file, Zoetis internal data, October 2016, Zoetis Inc.

EXPECTED PHENOTYPIC PREVALENCE IN FIRST LACTATION 40% 35% 30% 25% 20% 15% 10% 5% Metritis Mastitis Lameness 0% 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 Source: Data on file, Zoetis internal data, October 2016, Zoetis Inc.

SUMMARY AND CONCLUSIONS» We have made great progress genetically on Somatic Cell Score It has been and continues to be a good tool.» But, we have a new tool for genetic selection for lower incidences of mastitis and it is working out in the field.» To make the most progress in reducing mastitis it will take a combination of the best management practices along with a focus on the genetic side to give the best results.