Inbreeding Adjustments and Effect on Genetic Trend Estimates

1
Inbreeding Adjustments and Effect on Genetic
Trend Estimates
2
Adjustments Introduced Feb 2005

• Inbreeding depression
• Remove by regression in animal model
• Then include expected future inbreeding
• EFI .5 mean relationship to current cows
• Parity variance
• Heterogeneous variance refinement
• September 2004 test run included both

3
Inbreeding AdjustmentsVanRaden and Smith 1999
JDS 822771

• Subtract regression (b) from y
• YD0 y - m - p - c - b (Fcow)
• DYD0 and PTA0 reflect 0 inbreeding
• Add expected future inbreeding
• PTAEFI PTA0 b (EFI - EFIbase)
• PAEFI PA0 b (EFI - EFIbase)
• DYDEFI DYD0 b (EFI - EFIbase)

4
Effects of Inbreeding Adjustment

• Genetic evaluations of recent Holsteins
• Corr (PTA, PTAEFI) .998 for bulls
• Corr (PTA, PTAEFI) .993 for cows
• Genetic correlations of USA with other countries
  declined by .01 (Sep 1999 test run)
• Genetic trend
• Yield trends were 6 lower for PTAEFI
• Fertility, longevity trends 14-25 lower

5
Changes in Trends and Trend Testsfrom Inbreeding
Adjustments
Trait Trend Change SD/yr Trend Test Differences Trend Test Differences Trend Test Differences
Trait Trend Change SD/yr Test 1 Test 2 Test 3
Protein -.018 -.005 .000 .001
Fat -.015 -.001 .000 .000
SCS .002 .000 .000 .000
PL -.024 N/A .000 -.005
DPR -.009 .001 .001 .001
6
Parity Variance Adjustments

• Trend test 1 was .014 (fail), now .008 (pass) for
  Jerseys. Holsteins were OK.
• Deviations for yield are multiplied by 1.07 1.00
  .95 .90 .85 (parities 1-5)
• Genetic trend for protein decreased by .004
  genetic SD / year (2 change)
• Corr (official, adjusted PTA) .9998 for recent
  Holstein bulls

7
USA vs Non-USA Bull TrendsBefore (Aug) and After
(Sep) Adjustments
Genetic Trend (SD / yr) Genetic Trend (SD / yr) Genetic Trend (SD / yr) Genetic Trend (SD / yr) Ratio (Non- USA / USA) Ratio (Non- USA / USA)
USA bulls USA bulls Non-USA Non-USA Ratio (Non- USA / USA) Ratio (Non- USA / USA)
Scale Aug Sep Aug Sep Aug Sep
CAN .215 .209 .243 .234 1.13 1.12
DEU .178 .173 .214 .211 1.20 1.22
FRA .189 .184 .221 .218 1.17 1.18
NLD .182 .177 .219 .214 1.20 1.21
USA .231 .222 .239 .226 1.03 1.02
8
Effect on Top 100 Bulls

• Average number of USA bulls in top 100 across all
  27 protein scales
• Increase from 24.5 to 25.3 (Holstein)
• Increase from 53.8 to 56.3 (Jersey)
• USA genetic trend
• Decreased 6 (Holstein)
• Decreased 9 (Jersey)

9
Conclusions

• Boichard et al (1995) stated that biased genetic
  trend strongly disturbs international germplasm
  exchanges based on conversion formulas...
• MACE is robust to trend bias
• Trend tests should not be required