유전체 정보의 한우 육종 모델 적용에 관한 연구
= A Study on the Application of Genomic Information to Hanwoo (Korean Cattle) Breeding Model

 
저자 이승수
형태사항 x,89p. : 삽화 ; 26 cm.
일반주기 지도교수: 김내수.
학위논문사항 학위논문(박사)-- 충북대학교 대학원 : 축산학과 축산학전공 2013.2
KDC 527.4341 5
발행국 충청북도
언어 한국어
출판년 2013
소장기관 충북대학교 도서관

 
초록
The purpose of this study is to investigate the implementation of Hanwoo genetic evaluation as estimation of the accuracy of GEBV (Genomic Estimated Breeding Value) with GBLUP (Genomic Best Linear Unbiased Prediction). This study was conducted with 559 progeny-test bulls which were tested in Livestock Improvement Main Center from 45th, 46th, 48th and 49th progeny-test between 2009 and 2012. Pedigree data of interest were collected and 5 phenotype data (carcass weight, eye muscle area, backfat thickness, and marbling score) were recorded. Genotypes were analyzed with Illumina bovineHD BeadChip (777K SNPs).
 The results from this study are as follows ;

 1. The means and standard deviation of 552 Progeny-Test bulls were estimated 352.99±37.83㎏ for CW (Carcass Weight), 82.35±8.30㎠ for EMA (Eye Muscle Area), 7.99±3.08㎝ for BF (Back Fat) and 3.05±1.47 points for MS (Marbling Score), respectively. The effect of batch-number and the age at slaughter (month) affect significantly on carcass traits.

 2. Through quality control, out of 777,962 SNPs except sex chromosome SNPs, 180,083 SNPs were eliminated because those were over 20% of missing proportion and below 0.05 of minor allele frequence. 4,846 SNPs which had major homozygotes only and didn't have heterozygote were eliminated. 69,078 SNPs which showed extremely biased Hardy-Weinberg equilibrium were eliminated. Finally 578,489 SNPs were chosen and used to estimate.

 3. Imputation were conducted to estimate genotype which were omitted. The number of total SNP markers on used 552 individuals were 319,325,928 and imputation on 898,304 SNP markers were conducted. Estimated gene contents on missing value were between 1 for heterozygote and 2 for homozygote of second allele.

 4. Using least-square method, out of 578,489 SNPs, the number of significant SNP markers were estimated as 50,632 for CW, 50,435 for EMA, 67,213 for BF and 46,592 for MS, respectively.
 5. To estimate the accuracy of GEBV calculated by RR-BLUP (Ridge Regression Best Linear Unbiased Prediction), 10-fold cross validation method was used. Out of 552 animals which had genotypes and pedigree data, 90% (502 animals) were randomly selected as reference set to estimate the SNP marker effects and fixed effects in each traits. The rest 10% (50 animals) were used as validation to estimate the GEBV and this procedure repeated 10 times. The estimated relationship between to sets were quite high. The inbreeding coefficient based on genomic data was 0.4304 and the inbreeding coefficient based on pedigree data was 0.3926. The estimated accuracy of GEBV were 0.915 ~ 0.957 by each traits.

 6. To compare with GRM (Genomic Relationship Matrix) and NRM (Numerator Relationship Matrix) elements, All diagonal and off-diagonal 285,156 elements were investigated for distribution and regression. The estimated regression equation between NRM and GRM elements was Y = 0.8763X - 0.0227 (R2 : 0.7424).

 7. Estimated accuracies for 534 genotyped individuals using combined relationship matrix() which were consisted of GRM and NRM were increased at most 9.56%, 5.78%, 5.78% and 4.18% for CW, EMA, BF and MS, respectively. Estimated accuracies for un-genotyped individuals which were containing 3,674 co-ancestor using combined relationship matrix() were increased 13.54%, 6.50%, 6.50% and 4.31% for CW, EMA, BF and MS, respectively.

 
목차
Ⅰ. 서론
 
Ⅱ. 연구사
1. 육종가 추정 방법의 발전
2. 유전체 정보를 이용한 가축개량
3. 외국의 유전체 선발
4. 우리나라의 유전체 선발 연구

Ⅲ. 재료 및 방법
1. 공시 재료
2. 통계분석

Ⅳ. 결과 및 고찰
1. 표현형 자료 및 일반능력
2. 가계 및 유전체 자료를 이용한 혈통 분석
3. SNP 마커에 대한 정도 관리
4. 결측치에 대한 대치(Imputation)
5. SNP 마커의 효과 추정
6. 유전체 혈연 행렬
7. BLUP과 GBLUP으로 추정된 육종가의 정확도 비교
8. 고찰

Ⅴ. 적요

Ⅵ. 참고문헌