2016 AOA EPD Producer Report: Generation 9

Winter 2016
Dr. Mark Enns, Colorado State University

The new EPD for the 2016 genetic evaluation of the Huacaya and Suri databases were just completed. Expected progeny differences (EPD) and accuracies are now available for each trait. This year’s analysis included observations for at least one trait on 26,608 Huacaya and 6,497 Suri animals.

The following report summarizes the data used in the evaluation and the results of the EPD and accuracy calculations for each trait for both Huacaya and Suri. Percentile rankings by sex and possible change tables are included in this report. The animal rankings for each trait within the Huacaya and Suri populations can be determined from tables presented below. These rankings are split by sex (males and females) to give breeders a more precise idea of where their animals rank in the system. The percentile rankings for animals should be interpreted with care and in the context of the overall goals of the breeding program. Two primary considerations include:

  1. Males typically have many more progeny than females and therefore, on average have higher accuracies and a corresponding greater range in EPD for each trait.
  2. The value of fiber is a function of many traits and an animal excelling in one trait may not be superior in others—EPD and population rankings must be used in the context of the overall production system.

Data Summary

The number of observations for each trait used in the calculations is presented in Table 1 for both Huacaya and Suri. These tables include the average level of performance for each trait. The greatest increases in number of observations for the Huacaya population was for staple length and standard deviation of staple length, which increased by over 19.75%; with all traits increasing in number of observations by over 16%. In Suri, the greatest increase in observations was for staple length and standard deviation of staple length with over a 30% increase in observation numbers. All other trait observations grew by at least 25% excepting birth weight which increased by 14.5% in the Suri population.

The counts in Table 1 include the numbers of observations in the Alpaca Owners Association, Inc. database that were submitted and then available for use in the EPD calculations. As in previous analyses, some historical observations are outside of the range of allowable age at measure (age in days must be greater than 270 days), or there is not sufficient information to determine the age of the animal when the measurement was taken (i.e., missing birth or measurement/sample date). To be usable for calculating EPD, the age at which an animal is measured must be known. For the few animals where calculation of an age is not possible, the animal still received an EPD for the trait, but the EPD is based on the performance of relatives and does not include the animal’s own observation.

Table 1. Summary of observations for Huacaya and Suri by trait.
 

Huacaya

Suri

Trait Name

Count

Average

Count

Average

Average Fiber Diameter (microns-µ)

27,459

23.2

5,531

26.3

Standard Deviation (AFD; µ)

27,457

4.9

5,525

6.1

Spin Fineness (µ)

27,458

22.7

5,530

26.2

Percent of Fibers larger than 30 microns

27,458

12.3

5,455

24.6

Fleece Weight

26,094

6.1

5,600

5.4

Mean Curvature (deg/mm)

26,305

38.7

5,414

10.9

Standard Deviation of Curvature

26,303

23.3

5,414

13.7

Percent Medullation

12,283

15.8

2,274

15.4

Mean Staple Length (mm)

25,295

90.0

5,299

148.2

Birth Weight

13,370

16.5

3,261

17.4

 

To achieve the most accurate EPD possible, available performance data on all related individuals is used after being weighted by the animals’ genetic relationship. The calculation of the relationships amongst all animal in the evaluation population begins with a list of animals with performance information on any trait (e.g., fleece weight, fiber diameter). Animals on this list then have their pedigree traced back 4 generations and these “ancestors” are then included in the analysis. For those animals that do not have a 4-generation pedigree in the AOA registry (i.e., original importations and animals that are within 3 generations of an original import), pedigrees are constructed back to the original imports. Other than for ancestors of an animal with an observation, other animals are not included in the evaluation unless they have their own observation. For instance, progeny of animals with an observation are not included in the analysis unless those progeny have their own observation. Those “un-observed” progeny do not add information to the overall genetic evaluation. However a breeder could calculate a pedigree estimate EPD for those “un-observed” progeny by simply taking the average of their parents’ EPD. Collateral relatives (e.g., half-sibs, full-sibs, cousins) without an observation or without descendants with an observation, also add no additional information on genetic merit and are therefore not included in the analysis as well.

The constructed pedigree for Huacaya included 62,471 animals and for Suri included 14,364 animals. These animals and their relationships formed the basis for the EPD calculations.

Analysis Procedures

EPD and accuracy are both influenced by the degree of genetic influence on animal performance. This degree of genetic control, or heritability, is expressed as a decimal ranging from 0 to 1 (or as a percentage ranging from 0 to 100%), and varies for each trait. The heritability estimates used in this evaluation are shown in Table 2. As is clear, there is considerable genetic influence on the traits of interest in alpaca and many are greater than the heritability of traits of interest in other livestock species; therefore there is considerable opportunity for genetic improvement.

Table 2. Heritabilities used in the latest EPD calculations for Huacaya and Suri.

Trait

Huacaya

Suri

Average Fiber Diameter (AFD)

.52

.52

Standard Deviation of AFD

.52

.52

Spin Fineness

.52

.52

Percent of Fibers larger than 30 microns

.55

.52

Fleece Weight

.35

.32

Mean Curvature

.52

.51

Standard Deviation of Curvature

.55

.20

Percent Medullation

.54

.55

Mean Staple Length

.39

.15

Birth Weight

.50

.55

 

The magnitude of the heritability influences two practical aspects of EPDs and their calculation. First, the heritability determines the spread (i.e., minimum and maximum) of EPD for each trait—the greater the heritability, the greater the spread in EPD across all alpacas given the same amount of data (i.e., additional performance data also influences the spread of EPD in a population). Second, heritability influences the accuracy of the EPD—as heritability of a trait increases, the accuracy of the EPD increases given the same amount of performance information. At higher heritabilities, each observation (e.g., fiber diameter measure) is more closely related to the underlying genetics controlling expression of that trait and therefore, a single observation reveals much about the individual’s genetic merit. When heritability of a trait is low, a single observation on an individual reveals less about that animal’s genetic merit because environment has a larger influence on performance. For instance, the value of a single observation on fleece weight would result in a less accurate fleece weight EPD than would a single fiber diameter observation for the fiber diameter EPD. (The heritability of fleece weight is .35 (Huacaya) and the heritability of fiber diameter is .52.)

With the exception of birth weight, these EPD are calculated with multiple trait models that leverage information on genetically related traits in order to increase EPD accuracy. These genetic relationships enable the EPD system to use information on one trait to predict genetic merit in another because genes influencing one trait can also influence performance in the other traits of interest. The strength of that influence is reflected in the genetic correlation between two traits. For instance, the genetic correlation between fiber diameter and standard deviation of fiber diameter is .66 in Huacaya. A .66 genetic correlation indicates a strong tie between the two traits with many genes influencing fiber diameter also influencing standard deviation of fiber diameter.

Leveraging information in a multiple trait analysis results in more accurate EPD. The EPD from this analysis result from the analysis of 3-trait combinations with fleece weight as fleece weight had the greatest number of observations reported in both Huacaya and Suri in the original, developmental data set.

The birth weight analysis takes a slightly different approach. The difference, then, is that birth weight EPD are produced using only birth weight information in a single trait analysis. There are no correlated traits used in the calculation of birth weight EPD. The birth weight evaluation is similar to the other evaluations in that accounting for environmental (i.e., nongenetic) factors influencing performance is essential to producing more accurate EPD. For birth weight this is accomplished through “contemporary groups,” where a contemporary group is defined as all animals born on the same farm and in the same year. This combination is designed to account for climatic and nutritional variation from year to year on the same farm and for managerial differences between farms.

EPD Summary

Averages and ranges of EPD for each trait in both Huacaya and Suri are shown in Tables 3 through 6. For an animal to receive an EPD, it must have been within 4 generations of an animal with an observation and have an accuracy greater than 0 for fiber diameter. The summary statistics are split into different tables for males and females.

Table 3. Expected progeny differences and associated accuracies for all Huacaya males in the analysis.

 

Expected Progeny Differences

Accuracy

Trait

Average

Minimum

Maximum

Average

Maximum

Fiber Diameter (FD; µ)

-.3

-3.3

4.6

.17

.74

Standard Deviation of FD (µ)

-.1

-1.0

1.7

.17

.74

Spin Fineness (µ)

-.3

-3.4

4.6

.18

.75

Percent of Fibers > 30 microns

-1.4

-13.4

22.1

.18

.75

Mean Curvature (CURV)

.6

-6.6

10.9

.17

.74

Standard Deviation of CURV

.3

-3.7

5.8

.18

.75

Percent Medullation

-.3

-10.9

16.4

.09

.70

Staple Length (SL)

-.1

-13.5

12.1

.15

.71

Fleece Weight

.2

-1.2

1.9

.12

.68

Birth Weight

.0

-2.1

2.4

.13

.63

*Accuracies represent animals with an accuracy for fiber diameter greater than 0. As such, some animals may have an accuracy of 0 for other traits.

 
Table 4. Expected progeny differences and associated accuracies for all Huacaya females in the analysis.

 

Expected Progeny Differences

Accuracy

Trait

Average

Minimum

Maximum

Average

Maximum

Fiber Diameter (FD; µ)

-.2

-3.7

5.6

.15

.61

Standard Deviation of FD (µ)

-.1

-.9

1.8

.15

.61

Spin Fineness (µ)

-.2

-3.8

5.7

.16

.69

Percent of Fibers > 30 microns

-.9

-12.8

26.2

.16

.69

Mean Curvature (CURV)

.4

-7.3

10.9

.15

.59

Standard Deviation of CURV

.2

-3.5

5.3

.16

.63

Percent Medullation

-.2

-10.5

19.6

.08

.53

Staple Length (SL)

.0

-13.0

11.9

.13

.53

Fleece Weight

.1

-1.0

1.7

.10

.49

Birth Weight

.0

-2.7

2.4

.10

.39

*Accuracies represent animals with an accuracy for fiber diameter greater than 0. As such, some animals may have an accuracy of 0 for other traits.

 
Table 5. Expected progeny differences and associated accuracies for all Suri males in the analysis.

 

Expected Progeny Differences

Accuracy

Trait

Average

Minimum

Maximum

Average

Maximum

Fiber Diameter (FD)

-.2

-3.1

4.7

.17

.74

Standard Deviation of FD

-.1

-1.2

1.4

.17

.74

Spin Fineness (µ)

-.2

-3.1

3.7

.17

.75

Percent of Fibers > 30 microns

-.8

-13.3

17.4

.17

.75

Mean Curvature (CURV)

.0

-2.3

3.2

.17

.74

Standard Deviation of CURV

.0

-1.3

1.5

.10

.62

Percent Medullation1

-.6

-9.5

13.7

.09

.71

Staple Length

.4

-9.0

16.3

.09

.57

Fleece Weight

.1

-.8

1.4

.12

.66

Birth Weight

.0

-2.4

2.4

.13

.58

*Accuracies represent animals with an accuracy for fiber diameter greater than 0. As such, some animals may have an accuracy of 0 for other traits.

 
Table 6. Expected progeny differences and associated accuracies for all Suri females in the analysis.

 

Expected Progeny Differences

Accuracy

Trait

Average

Minimum

Maximum

Average

Maximum

Fiber Diameter (FD)

-.1

-3.2

5.9

.16

.58

Standard Deviation of FD

.0

-1.0

1.7

.16

.55

Spin Fineness (µ)

-.1

-3.3

5.8

.16

.64

Percent of Fibers > 30 microns

-.5

-14.3

23.2

.16

.62

Mean Curvature (CURV)

.0

-2.8

7.0

.16

.63

Standard Deviation of CURV

.0

-1.5

2.3

.09

.41

Percent Medullation1

-.4

-10.5

20.4

.08

.59

Staple Length

.3

-9.7

20.4

.08

.34

Fleece Weight

.1

-.9

1.4

.10

.42

Birth Weight

.0

-2.0

2.3

.09

.37

*Accuracies represent animals with an accuracy for fiber diameter greater than 0. As such, some animals may have an accuracy of 0 for other traits.

 

Accuracy values provided with each EPD offer an easy method for evaluating confidence in that prediction (i.e., EPD) with accuracies closer to 1 yielding more confidence than EPD with lower accuracies. No matter the accuracy, however, the EPD for a trait is a consistently better prediction of an animal’s genetic merit than its own performance alone because EPD are based on considerably more data than an observation from only the individual. EPD take into account data from ancestors, collateral relatives (e.g., half sibs, full sibs), the individual itself, and progeny (if available); adjust for differences in age of the animal at measure; and account for nutritional and climatic differences across herds with appropriately designated contemporary groups.

As data accumulates on an individual and its relatives, the accuracy of the EPD for that individual increases. As an example, for a trait that is 50% heritable, with only a single observation per individual, no inbreeding, no other information on relatives, and an infinitely large contemporary group (this does not occur other than theoretically), the accuracy of the EPD should be just over .28. An advantage of the statistical methodology used to calculate these EPD is that animals do not need to compete in infinitely large contemporary groups. The system adjusts accuracies up or down accordingly based on the number of animals in the contemporary group with larger contemporary groups having greater influence on the EPD. Given the same number of animals in a contemporary group, traits with lower heritability than .5 will result in a lower accuracy associated with a single observation.

Another perspective on accuracy is available through the use of possible change values for each trait (Tables 7 and 8). The possible change values can be used to construct a confidence interval around an EPD and within that confidence range we are confident (within a given probability) that the animal’s true genetic merit resides. To illustrate, let’s assume a particular Huacaya animal’s fiber diameter EPD is -1.2 with an accuracy of .4. That EPD (-1.2) plus/minus possible change will produce a range within which we are 68% confident the animal’s true merit for fiber diameter lies. With a .4 accuracy, the possible change value for fiber diameter in this example is .70 (Table 7) and -1.2 minus .70 and -1.2 plus .70 results in a confidence range of -1.9 (-1.2-.70) to -.40 (-1.2+.70). We are 68% confident the animal’s true genetic merit lies in that range. As is apparent from the tables and from intuition, as we gather more data on an animal and its relatives, we become more confident in our EPD—accuracy increases and possible change values decrease (Tables 7 and 8). Accordingly, the confidence range would be narrower as our estimate is more precise.

Table 7. Possible change values for various accuracy levels by trait in the Huacaya population.

Accuracy

FD

SD of FD

SPIN

Percent of Fibers > 30μ

CURV

SD of CURV

MED

SL

FW

BW

0.0

1.17

0.35

1.17

5.37

2.48

1.27

5.24

4.21

0.47

0.95

0.1

1.06

0.32

1.05

4.83

2.23

1.14

4.72

3.79

0.42

0.86

0.2

0.94

0.28

0.93

4.30

1.98

1.02

4.19

3.37

0.38

0.76

0.3

0.82

0.25

0.82

3.76

1.74

0.89

3.67

2.95

0.33

0.67

0.4

0.70

0.21

0.70

3.22

1.49

0.76

3.14

2.53

0.28

0.57

0.5

0.59

0.18

0.58

2.69

1.24

0.63

2.62

2.10

0.23

0.48

0.6

0.47

0.14

0.47

2.15

0.99

0.51

2.10

1.68

0.19

0.38

0.7

0.35

0.11

0.35

1.61

0.74

0.38

1.57

1.26

0.14

0.29

0.8

0.23

0.07

0.23

1.07

0.50

0.25

1.05

0.84

0.09

0.19

0.9

0.12

0.04

0.12

0.54

0.25

0.13

0.52

0.42

0.05

0.10

1.0

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

AFD=Fiber Diameter, SPIN=Spin Fineness, CURV=Mean Curvature, MED=Percent medullation, SL=Staple length, FW=Fleece weight, SD = standard deviation, BW=Birth weight.

 
Table 8. Possible change values for various accuracy levels by trait in the Suri population.

Accuracy

FD

SD of FD

SPIN

Percent of Fibers > 30μ

CURV

SD of CURV

MED

SL

FW

BW

0.0

1.45

0.49

1.48

6.81

0.99

0.82

5.54

6.64

0.45

0.95

0.1

1.30

0.44

1.33

6.13

0.89

0.74

4.99

5.98

0.40

0.86

0.2

1.16

0.39

1.19

5.44

0.79

0.66

4.43

5.31

0.36

0.76

0.3

1.01

0.34

1.04

4.76

0.69

0.57

3.88

4.65

0.31

0.67

0.4

0.87

0.29

0.89

4.08

0.59

0.49

3.33

3.99

0.27

0.57

0.5

0.72

0.25

0.74

3.40

0.49

0.41

2.77

3.32

0.22

0.48

0.6

0.58

0.20

0.59

2.72

0.40

0.33

2.22

2.66

0.18

0.38

0.7

0.43

0.15

0.44

2.04

0.30

0.25

1.66

1.99

0.13

0.29

0.8

0.29

0.10

0.30

1.36

0.20

0.16

1.11

1.33

0.09

0.19

0.9

0.14

0.05

0.15

0.68

0.10

0.08

0.55

0.66

0.04

0.10

1.0

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

AFD=Fiber Diameter, SPIN=Spin Fineness, CURV=Mean Curvature, MED=Percent medullation, SL=Staple length, FW=Fleece weight, SD = standard deviation, BW=Birth weight.

 

To make it easier for breeders to determine how their animals rank relative to the overall alpaca population, percentile tables for Huacaya and Suri are shown in Tables 9 through 12. Percentiles are determined by sex—tables split into male and female percentiles for Huacaya and Suri. These show the relative ranking of an individual in comparison to the population of animals of the same sex (those with published EPD). For instance, if a Huacaya male had a fiber diameter EPD of -1.60 it would be in the top 5% of all Huacaya males for fiber diameter.

No percentile rankings are provided for traits for which there are no clear superior performance endpoints (high or low) or for which there are likely intermediate optimums. In the case of traits with an intermediate optimum, the best animal is not at either extreme, high or low. A good example is birth weight where the best animal is neither too light nor too heavy at birth but rather has an intermediate birth weight. In the AOA analysis, these traits include mean curvature, standard deviation of mean curvature, and birth weight for Suri; and standard deviation of mean curvature and birth weight for Huacaya.

Table 9. Percentile rankings in Huacaya males for each trait*.

Percentile

FD

SDFD

SPIN

PERC

CURV

MED

SL

FW

1

-2.12

-0.60

-2.17

-7.79

5.82

-5.34

5.36

0.99

2

-1.93

-0.55

-1.96

-7.03

5.08

-4.59

4.44

0.88

3

-1.79

-0.51

-1.84

-6.59

4.63

-4.15

3.89

0.79

4

-1.69

-0.48

-1.73

-6.27

4.25

-3.84

3.55

0.74

5

-1.60

-0.45

-1.64

-6.03

3.95

-3.54

3.23

0.69

10

-1.32

-0.37

-1.34

-5.15

3.02

-2.69

2.27

0.56

15

-1.11

-0.31

-1.13

-4.48

2.40

-2.13

1.74

0.47

20

-0.94

-0.26

-0.96

-3.91

1.96

-1.72

1.34

0.41

25

-0.80

-0.22

-0.81

-3.46

1.58

-1.40

1.05

0.36

30

-0.68

-0.19

-0.69

-3.06

1.26

-1.12

0.78

0.31

35

-0.56

-0.16

-0.57

-2.65

0.99

-0.89

0.56

0.26

40

-0.47

-0.13

-0.47

-2.24

0.74

-0.67

0.34

0.22

45

-0.37

-0.10

-0.38

-1.87

0.53

-0.47

0.14

0.18

50

-0.28

-0.08

-0.28

-1.51

0.33

-0.28

-0.03

0.14

60

-0.12

-0.03

-0.12

-0.81

-0.01

0.08

-0.42

0.07

70

0.05

0.02

0.05

-0.06

-0.35

0.47

-0.88

0.01

80

0.25

0.09

0.25

0.80

-0.75

1.00

-1.47

-0.05

90

0.54

0.18

0.55

2.29

-1.35

1.89

-2.47

-0.14

*Where FD=Fiber diameter; SDFD=Standard deviation of fiber diameter; SPIN=Spin fineness; PERC=Percent of Fibers >30 microns; CURV=Mean curvature; MED=Percent medullation; SL=Staple length; FW=Fleece weight.

 
Table 10. Percentile rankings in Huacaya females for each trait*.

Percentile

FD

SDFD

SPIN

PERC

CURV

MED

SL

FW

1

-2.09

-0.58

-2.13

-7.57

5.51

-5.02

5.01

0.89

2

-1.84

-0.52

-1.88

-6.82

4.71

-4.35

4.12

0.78

3

-1.69

-0.47

-1.73

-6.36

4.17

-3.89

3.62

0.71

4

-1.58

-0.44

-1.61

-6.00

3.80

-3.51

3.25

0.66

5

-1.49

-0.42

-1.52

-5.73

3.50

-3.24

2.99

0.62

10

-1.16

-0.33

-1.18

-4.74

2.57

-2.37

2.09

0.50

15

-0.95

-0.27

-0.96

-4.01

1.98

-1.82

1.61

0.41

20

-0.78

-0.22

-0.79

-3.42

1.53

-1.43

1.25

0.34

25

-0.64

-0.18

-0.64

-2.90

1.18

-1.12

0.95

0.28

30

-0.52

-0.15

-0.52

-2.45

0.90

-0.86

0.72

0.23

35

-0.41

-0.12

-0.41

-2.00

0.66

-0.63

0.52

0.18

40

-0.31

-0.09

-0.32

-1.61

0.46

-0.44

0.34

0.14

45

-0.23

-0.07

-0.23

-1.24

0.27

-0.27

0.16

0.11

50

-0.15

-0.04

-0.15

-0.89

0.12

-0.12

0.01

0.07

60

-0.01

0.00

-0.01

-0.24

-0.14

0.12

-0.31

0.01

70

0.12

0.04

0.13

0.38

-0.43

0.47

-0.71

-0.03

80

0.30

0.10

0.32

1.26

-0.81

0.97

-1.23

-0.08

90

0.60

0.19

0.61

2.72

-1.39

1.88

-2.15

-0.15

*Where FD=Fiber diameter; SDFD=Standard deviation of fiber diameter; SPIN=Spin fineness; PERC=Percent of Fibers >30 microns; CURV=Mean curvature; MED=Percent medullation; SL=Staple length; FW=Fleece weight.

 
Table 11. Percentile rankings in Suri males for each trait*.

Percentile

FD

SDFD

SPIN

PERC

MED

SL

FW

1

-2.10

-0.71

-2.14

-9.39

-5.94

8.77

0.78

2

-1.84

-0.62

-1.89

-8.51

-5.14

6.97

0.69

3

-1.71

-0.58

-1.75

-7.79

-4.72

6.04

0.62

4

-1.61

-0.55

-1.65

-7.31

-4.32

5.37

0.56

5

-1.52

-0.52

-1.55

-7.00

-4.01

5.05

0.53

10

-1.19

-0.41

-1.25

-5.60

-3.07

3.47

0.42

15

-0.96

-0.34

-1.01

-4.65

-2.51

2.64

0.35

20

-0.80

-0.28

-0.84

-3.85

-2.05

2.09

0.30

25

-0.65

-0.24

-0.71

-3.22

-1.69

1.60

0.25

30

-0.54

-0.19

-0.58

-2.68

-1.38

1.19

0.21

35

-0.43

-0.16

-0.47

-2.16

-1.11

0.83

0.17

40

-0.33

-0.12

-0.36

-1.67

-0.85

0.54

0.13

45

-0.24

-0.09

-0.26

-1.21

-0.61

0.29

0.10

50

-0.16

-0.06

-0.17

-0.73

-0.38

0.06

0.07

60

0.03

0.00

0.00

0.00

0.00

-0.33

0.02

70

0.20

0.05

0.17

0.86

0.34

-0.78

-0.02

80

0.43

0.13

0.38

1.92

0.82

-1.38

-0.07

90

0.85

0.26

0.75

3.72

1.56

-2.35

-0.14

*Where FD=Fiber diameter; SDFD=Standard deviation of fiber diameter; SPIN=Spin fineness; PERC=Percent of Fibers >30 microns; MED=Percent medullation; SL=Staple length; FW=Fleece weight.

 
Table 12. Percentile rankings in Suri females for each trait*.

Percentile

FD

SDFD

SPIN

PERC

MED

SL

FW

1

-2.04

-0.69

-2.09

-9.37

-5.95

7.86

0.72

2

-1.80

-0.61

-1.84

-8.24

-5.10

6.24

0.61

3

-1.65

-0.56

-1.69

-7.58

-4.54

5.39

0.55

4

-1.52

-0.52

-1.57

-7.05

-4.15

4.87

0.51

5

-1.41

-0.49

-1.46

-6.61

-3.89

4.41

0.48

10

-1.08

-0.38

-1.12

-5.15

-2.96

3.09

0.38

15

-0.87

-0.31

-0.91

-4.18

-2.29

2.34

0.31

20

-0.71

-0.25

-0.74

-3.42

-1.78

1.75

0.25

25

-0.57

-0.20

-0.60

-2.74

-1.40

1.29

0.21

30

-0.44

-0.16

-0.47

-2.15

-1.07

0.94

0.17

35

-0.33

-0.12

-0.36

-1.62

-0.77

0.66

0.13

40

-0.24

-0.09

-0.26

-1.17

-0.51

0.44

0.10

45

-0.15

-0.07

-0.17

-0.76

-0.32

0.23

0.07

50

-0.08

-0.04

-0.09

-0.36

-0.14

0.05

0.04

60

0.06

0.01

0.04

0.20

0.08

-0.30

0.01

70

0.23

0.06

0.19

0.96

0.39

-0.73

-0.03

80

0.45

0.13

0.40

1.99

0.84

-1.34

-0.07

90

0.86

0.26

0.76

3.77

1.64

-2.25

-0.15

*Where FD=Fiber diameter; SDFD=Standard deviation of fiber diameter; SPIN=Spin fineness; PERC=Percent of Fibers >30 microns; MED=Percent medullation; SL=Staple length; FW=Fleece weight.

 

After each EPD analysis, diagnostic tests are performed to identify any unusual changes in EPD and/or any potential data integrity issues. One of these tests is to calculate a rank correlation between EPD of animals in the previous analysis with new EPD on those same animals in the latest analysis. The rank correlation evaluates reranking of animals in the latest analysis with a rank correlation of 1.0 indicating no reranking of animals while a rank correlation of 0 indicating no relationship between EPD from the last and current analyses. Both extremes, a zero and a 1.0 correlation, are undesirable. The former would indicate that data in the previous analysis was completely overwhelmed by new data—a situation that should never occur. The latter, a 1.0 correlation, would indicate the additional data did not add any new information or accuracy to animals in the previous analysis. Obviously this is undesirable as well. As such, correlations closer to 1 but not 1 are desired. For Huacayas the lowest correlations for all animals between EPD from the previous 2015 evaluation and the current analysis was a .97 for staple length and for standard deviation of mean curvature. The highest were above .98 for fiber diameter, standard deviation of fiber diameter, spin fineness, percent fibers greater than 30 microns, mean curvature, percent medullation and fleece weight. In Suris, the lowest correlation was for staple length (.935) with the remaining trait correlations all higher than that level.

As with the 2015 analysis, the base for the genetic evaluation has been fixed. The term “base” refers to a group of animals from which all genetic comparisons begin. The base population is set as all animals with no parentage information. Likely these are the animals from which the US population arose (i.e., original imports). As historical information has been added to the database more animals from the original importations have been included in the EPD calculations resulting in a scenario where this group of animals has become more consistent from one analysis to the next.

Conclusions

The performance database continues to expand with additional data. Accordingly, the EPD are becoming more reliable with maximum accuracies in the population continuing to increase. With the increase in data, evaluations have also become more stable with the correlations of EPD from one analysis to the next continuing to increase.