Abstr 422 Harnessing automatic data collection to enhance

Abstr. 422 Harnessing automatic data collection to enhance genetic improvement programs G. R. Wiggans, [email protected]. arsusda. gov Animal Improvement Programs Laboratory, Agricultural Research Service, USDA, Beltsville, MD M. A. Faust ABS Global, Inc. , Deforest, WI F. Miglior Agriculture and Agri-Food Canada, Sherbrooke, QC, Canada Canadian Dairy Network, Guelph, ON, Canada ADSA 2007 (1) G. R. Wiggans 2007

Questions l What should we be thinking about to prepare for the future? l How can we best exploit technology that is/will be available? l How can we minimize the impact of negative trends? ADSA 2007 (2) G. R. Wiggans 2007

History of innovation 1950 s Computerization 1960 s Laboratories for component testing 1970 s Farm to mainframe connection for input and reports 1980 s Electronic data transfer from farms and laboratories; on-farm data entry 1990 s Robotic/voluntary milking systems 2000 s Handheld devices for data collection and access; RFID ADSA 2007 (3) G. R. Wiggans 2007

Automatic data collection l l ADSA 2007 (4) Continued adoption of technology Includes equipment or procedures that aid data collection G. R. Wiggans 2007

Electronic milk meters l Currently supply 7% of data l Can provide −Total yield − Milking speed − Milk conductivity l May provide − Progesterone levels − Milk temperature − Component concentration l RFID may improve reliability of cow ID associated with meter data ADSA 2007 (5) G. R. Wiggans 2007

Voluntary milking systems l Also known as robotic parlors l Most common in Europe l Depend heavily on automatic data collection l Require adaptation by DHI to be included l May provide data not available elsewhere ADSA 2007 (6) G. R. Wiggans 2007

Other data collection devices l l Handheld computers to record health l Activity monitors l ADSA 2007 (7) Electronic scales Weather stations G. R. Wiggans 2007

New traits l l Heat tolerance l ADSA 2007 (8) Diagnostic test results Lameness G. R. Wiggans 2007

Limitations l l Equipment failure l Staff capability l ADSA 2007 (9) Electronic ID read errors Cost G. R. Wiggans 2007

Why farms should invest in automatic data collection l Better management w More accurate data w More characteristics w Greater quality control l Food quality assurance and traceability l Help genetic improvement? ADSA 2007 (10) G. R. Wiggans 2007

Trends l More traits recorded l Larger herds l Improved equipment for electronic recording l Increased worldwide competition among AI organizations w Demand for increased data accuracy and comprehensiveness, especially for traits with low heritability ADSA 2007 (11) G. R. Wiggans 2007

Needs of genetic improvement program l l Maintained or improved data quality l ADSA 2007 (12) Continued participation Adaptation to change G. R. Wiggans 2007

Benefits to genetic improvement from automatic data collection l l Reduced cost l ADSA 2007 (13) Improved accuracy More traits G. R. Wiggans 2007

Tradeoffs in adding traits l l Recording errors l Difficulty in estimating economic value l ADSA 2007 (14) Low heritability Dissipation of selection differential G. R. Wiggans 2007

Why more traits? l Goal of a profitable cow l Selection index w Evaluations weighted by economic contribution l More precise measurement of profitability w More accurate profit tracking w More accurate selection ADSA 2007 (15) G. R. Wiggans 2007

How to connect genetic improvement to on-farm data l Provide value w Genetic evaluations w Data backup w Data quality control l Compensate for data as a dairy product (like milk) l Promote connection ease and security ADSA 2007 (16) G. R. Wiggans 2007

On-farm software l Must be maintained l Support w Extremely labor intensive w Expensive if many platforms l Central control of updates attractive l Dedicated uniform hardware? ADSA 2007 (17) G. R. Wiggans 2007

Systems for farms to provide data l Current system w AI organizations pay for progeny-test daughters w Bundled with DHI program l System managed by AI organizations w AI organizations connect to on-farm computers w Data quality monitored by AI organizations l Farm as data vendor w Farm markets data to AI organizations w Compensation based on quality ADSA 2007 (18) G. R. Wiggans 2007

Who is in charge? l AI organizations w Establish data connection with progeny-test herds l DRPC w Offer test plans that provide desired data l Farm w Market data based on quality l Cooperation w Establish mechanism for equitable resolution of competing interests ADSA 2007 (19) G. R. Wiggans 2007

Measures of data quality l Consistency w Milk weights vs. milk shipped w Calving, progeny birth, breeding, dry dates l ID accuracy from parentage verification l Electronic ID w Protocols to detect misreads w Portion of duplicate or missing cows l Within-herd heritability ADSA 2007 (20) G. R. Wiggans 2007

Management vs. genetic improvement l Large-herd management based on cow groups l Selection based on evaluation of individuals l Genetic improvement needs data from individuals ADSA 2007 (21) G. R. Wiggans 2007

Where genetic improvement needs to go beyond herd management needs l Accurate ID l Access to all data w Allow efficient research and development of new trait evaluations l Sufficient incentive for herds to participate ADSA 2007 (22) G. R. Wiggans 2007

Herd of the future l Every milking recorded and components determined l All calves genotyped w Parentage verification w Genomic-based evaluation l Data delivered to evaluation center daily ADSA 2007 (23) G. R. Wiggans 2007

Hurdles for SNP l Benefits justified by cost l Convenient DNA collection and accurate labels l Timely and adequately accurate genomic prediction l For parentage verification/discovery, genotypes from same SNPs required for potential parents ADSA 2007 (24) G. R. Wiggans 2007

Evaluations on demand l l ADSA 2007 (25) Estimates of SNP effects updated several times each year Genomic prediction calculated as soon as genotype available G. R. Wiggans 2007

Best practice l Collection of accurate data for all relevant traits l Seamless transfer to evaluation center l Evaluations calculated with test-day model and including genomic data l Results available as needed ADSA 2007 (26) G. R. Wiggans 2007

What to expect from automatic data collection ADSA 2007 (27) G. R. Wiggans 2007

Incentives l Quality data have value l Computer capacity on farm minimizes need for central computing l Economic incentive required for dairies to contribute data to national evaluations l Appropriate to have incentive based on data quality ADSA 2007 (28) G. R. Wiggans 2007

Benefits to herds l Improved management information l Incentives from AI organizations for providing data l Improved pedigree accuracy from parentage validation ADSA 2007 (29) G. R. Wiggans 2007

Impact on national evaluations l More traits w Body condition score based on electronic scales w Mobility l Higher quality w Electronic recording and monitoring l Lower cost w Less labor required ADSA 2007 (30) G. R. Wiggans 2007

Lower cost l l ADSA 2007 (31) Technician cost on test-day eliminated On-farm component determination G. R. Wiggans 2007

Conclusions l l ADSA 2007 (32) Automated data collection w Growing w Can improve data quality Genetic improvement programs w More traits w Better inputs w Tighter connection to sources G. R. Wiggans 2007