1	USDA Genetic Evaluation Program for Dairy Goats
2	Why Genetic Evaluations? A valuable tool for genetic selection Allows for comparison of animals in different environments Can include all of the information available for each animal Greatest impact on progress is from selection for males
3	Why Genetic Selection? Genetic selection can improve fitness, utility, and profitability Females must be bred to provide replacements and initiate milk production Mate selection is an opportunity to make genetic change
4	Selection is a Continuous Process Decisions Which females to breed Which males to use Which specific matings to make Which progeny to raise Which females to keep and breed Goals Improve production and efficiency Avoiding inbreeding Correct faults
5	Genetic Improvement Program Phenotype = Genotype + Environment Genetic improvement programs only change genotype Rate of genetic improvement determined by: Generation interval Selection intensity Heritability Heritability is the portion of total variation due to genetics
6	Steps in Genetic Evaluation Define a breeding goal Measure traits related to the goal Record pedigree to allow detection of relationships across generations Identify non-genetic factors that affect records and could bias evaluations Make adjustments Include in the model Define an evaluation model
7	Examples of Breeding Goals Increased milk, fat, or protein yield Increased longevity Optimal number of kids born Improved conformation score (overall and linear) Increased profitability
8	Examples of Non-genetic Factors Age Lactation Season Litter size Milking frequency Herd
9	Data Flow
10	Does on Test at Last Test in 2005 By Processing Center
11	Data Validation Incoming data is checked against database for verification Birth date is checked against kidding date Sire and dam are checked against breeding records and ADGA Cross-references are assigned when identification changes
12	Data Validation (Cont.) Cross-references are determined based on control number Abnormal yields are detected and reported to DRPC Test dates and testing characteristics are compared with herd data
13	Alpine Milk Production Lactation Curve
14	Alpine Fat Percentage Lactation Curve
15	Alpine Protein Percentage Lactation Curve
16	Alpine and Nubian Milk Production Second Lactation
17	Nubian Fat and Protein Percentage Second Lactation
18	Evaluation Calculation Goal Predict productivity of progeny Method Separate genetic component from other factors influencing evaluated traits All relationships are considered Bucks receive evaluations from the records on their female relatives
19	Evaluation model An equation that indicates what factors contribute to an observation Separates the genetic component from other factors Solutions used to predict the genetic potential of progeny
20	Yield Model: y = hys + hs + pe + a + e y = yield of milk, fat, or protein during a lactation hys = herd-year-season Environmental effects common to lactations in the same season, within a herd hs = herd-sire Effects common to daughters of the same sire, within a herd pe = permanent environment Non-genetic effect common to all of a doe’s lactations a = animal genetic effect (breeding value) e = unexplained residual
21	Indexes An index combines evaluations for a group of traits based on their contribution to a selection goal Milk-Fat-Protein Dollars Combines yield evaluations into a single number MFP$ = 0.01(PTA_Milk) + 1.15(PTA_Fat) + 2.55(PTA_Protein)
22	Type Traits Describe physical characteristics of animal Final Score (overall assessment) Scored 50-99 Linear traits (13 defined traits) Scored 1-50
23	Type Evaluation Model MODEL: y = h + a + p + e y = Adjusted type record h = Herd appraisal date a = Animal genetic effect (breeding value) p = Permanent environment - Effect common to all a doe's lactations that is not genetic e = Unexplained residual Multi-trait - Scores of one trait affect evaluations of other traits.
24	Type Trait Genetic Correlations
25	Combining type and production Production-Type index (PTI) Combines yield and type evaluations into a single value There are 2 versions: PTI 2:1, weights 2 production : 1 type PTI 1:2, weights 2 type : 1 production
26	How Accurate are Evaluations? Reliability measures the amount of information contributing to an evaluation Increases as daughters are added (at decreasing rate) Also affected by: Number of contemporaries Reliability of parents’ evaluations Heritability
27	Accuracy of Evaluations Does kidding in same season More records ® better estimate of herd-year-season (hys) effect Bucks with daughters having records in same hys More direct comparisons ® better ranking of bucks Number of lactation records Number of daughters Completeness of pedigree data
28	Methods of Expressing Evaluations Estimated breeding value (EBV) Animal’s own genetic value Predicted transmitting ability (PTA) ½ EBV Expected contribution to progeny
29	Heritability Portion of total variation due to genetics Milk, Fat, Protein: 25% Range for Type: 19% (r. udder arch) — 52% (stature)
30	USDA Dairy Goat Evaluations Evaluations for milk, fat, protein, and type Yield evaluations in July Type evaluations in November Evaluations provided to ADGA, DRPC, and public via the Internet (aipl.arsusda.gov)
31	What Do the Numbers Mean? Evaluations are predictions The true value is unknown The predictions rank animals relative to one another using a defined base The base is the zero- or center-point for evaluations For example: the performance of animals born in a given year
32	Trend in Breeding Value for Milk
33	Ways to Increase Rate of Improvement Use artificial insemination (AI) to use better males in more herds Identify promising young males for progeny testing (PT) Use on a representative group of does and observe the actual success of progeny Focus on larger herds to improve accuracy
34	Factors Affecting Value of Data Completeness of ID and parentage reporting Years herd on test Size of herd Frequency of testing and component determination
35	Why Evaluations Go Wrong Important factors ignored Litter size Milking Frequency Preferential treatment Unlucky Current data not representative of future data Traits with low heritability require large numbers to be accurate Recording errors Wrong daughters assigned to a sire
36	Dairy Cattle Program for Genetic Improvement Artificial insemination (AI) Allows for many progeny from superior males Allows semen to be used in geographically diverse locations Progeny testing (PT) Use young males to get a representative group of daughters Wait until those daughters are milking Based on the evaluations, return the best males to heavy use
37	Dairy Cattle Program for Genetic Improvement (Cont.) Pre-select only promising bulls for PT Select only the best of the PT bulls for widespread use Only about 1 in 10 PT bulls enter active service Remove bulls from active service as better new bulls become available Bulls remain active only a few years
38	Alternative to Waiting for PT Use young bucks for most breedings Replace bucks quickly Bank semen of young bucks Use frozen semen from superior proven bucks as sires of next generation of young bucks
39	Recent Changes to System Web query for accessing data by animal name Yield data since 1998 extracted from the master file each run Incorporates corrections, deletions, and ID changes Standardized yields back to 1974 available
40	Recent Changes to System (Cont.) Added Breed codes CC – Sable ND – Nigerian Dwarf ID simplified by removing G and 18 prefixes when not required for uniqueness More complete breeding information stored
41	Possible Enhancements Add evaluations for more traits Productive Life Somatic Cell Score Daughter Pregnancy Rate Switch to test day model Provides better accounting for environment Accounts for genetic differences in shape of lactation curve
42	Future DNA analysis Parentage verification Genetic evaluation Genomic information may enable reasonably accurate evaluation at birth National Animal Identification System (NAIS) May cause changes in ID
43	Genomic Data Single Nucleotide Polymorphisms (SNP) Large number of markers with 2 alleles Tags segments of chromosomes Parentage verification Marker alleles must match those of a parent Often can infer unknown parent ID EBV calculated for chromosome segments Sum the value of segments to approximate evaluation Accuracy may approach progeny test
44	Conclusions Genetic evaluations are available for type and production Traits can be improved through selection Rate of improvement increases with accuracy of evaluations AI enables widespread use of superior bucks and enables PT bucks to be used across herds
45	Conclusions (cont.) Genetic evaluations improve selection accuracy Accurate evaluations also require adequate data and an appropriate model Evaluations are based on comparisons Differences for non-genetic reasons must be removed DNA technology is of great interest Still requires reliable evaluations
46	AIPL web services http://aipl.arsusda.gov/query/public/tdb.shtml#GoatsTBL Queries provide display of: Pedigree information Yield records Herd test characteristics Genetic evaluations of does & bucks Yield Type Access information using: ID number Animal name Herd code