wombat으로 fixed regression model 풀기

wombat으로 fixed regression model 풀기

 

예제 자료

 

R.A. Mrode, Linear Models for the Prediction of Animal Breeding Values. 2nd. Edition

 

Page 137. Example 7.1

 

자료입력

 

4 4 1 4 17.0

4 4 2 38 18.6

4 4 3 72 24.0

4 4 4 106 20.0

4 4 5 140 20.0

4 4 6 174 15.6

4 4 7 208 16.0

4 4 8 242 13.0

4 4 9 276 8.2

4 4 10 310 8.0

5 5 1 4 23.0

5 5 2 38 21.0

5 5 3 72 18.0

5 5 4 106 17.0

5 5 5 140 16.2

5 5 6 174 14.0

5 5 7 208 14.2

5 5 8 242 13.4

5 5 9 276 11.8

5 5 10 310 11.4

6 6 6 4 10.4

6 6 7 38 12.3

6 6 8 72 13.2

6 6 9 106 11.6

6 6 10 140 8.4

7 7 4 4 22.8

7 7 5 38 22.4

7 7 6 72 21.4

7 7 7 106 18.8

7 7 8 140 18.3

7 7 9 174 16.2

7 7 10 208 15.0

8 8 1 4 22.2

8 8 2 38 20.0

8 8 3 72 21.0

8 8 4 106 23.0

8 8 5 140 16.8

8 8 6 174 11.0

8 8 7 208 13.0

8 8 8 242 17.0

8 8 9 276 13.0

8 8 10 310 12.6

 

개체, 개체(영구환경효과), HTD(herd-test-day), DIM(days in milk), test day fat yield

 

위 자료를 data.txt로 저장

 

혈통입력

 

1 0 0

2 0 0

3 0 0

4 1 2

5 3 2

6 1 5

7 3 4

8 1 7

 

개체, 아비, 어미

 

위 자료를 pedi.txt로 저장

 

파라미터 파일 작성

 

# run option - 육종가를 구할 때

# RnSoln_xxx.dat 출력 파일 확인

RUNOP -v --solvit

 

# run option - 육종가와 SEP(standard error of prediction)

# reliability(r2) = 1- SEP^2 / sigma_a^2

# RnSoln_xxx.dat 출력 파일 확인

#RUNOP -v --blup

 

# run option - 좋은 초기값으로 분산성분을 추정할 때,

#RUNOP -v --good

 

# run option - 나쁜 초기값일 때 분산성분을 추정할 때,

#RUNOP -v --bad

 

# 요약 출력파일에 출력할 내용

COMMENT Fixed Regression Model from Mrode, 2nd Edition, Example 7.1

 

# Analysis Type

ANALYSIS UNI

 

# 혈통 파일 이름

# SumPedigree.out 확인

PED pedi.txt

 

# 자료 파일 이름

DATA data.txt

animal

pe_ani 99

htd 10

dim

milk_yld

END DATA

 

# Model of analysis

# SumModel.out 확인

MODEL

TR milk_yld

FIX htd

COV dim(4,LEG)

RAN animal NRM

RAN pe_ani IDE

END MODEL

 

# 분산 성분

VAR animal 1

5.521

 

VAR pe_ani 1

8.47

 

VAR error 1

3.71

 

# SumModel.out 확인

# SumEstimates.out 확인

# SumPedigree.out 확인

# FixSolutions.out 확인

# RnSoln_xxx.dat 확인

 

위 파라미터 파일을 wombat.par로 저장

 

실행

 

위 세 파일을 한 폴더에 넣고 다음과 같이 실행

결과 확인

 

SumModel.out

 

======= Version 19-05-2012 ======================================= **KM** ====

 

Program WOMBAT : Summary of information from Set-up step

==============================================================================

 

Fixed Regression Model from Mrode, 2nd Edition, Example 7.1

 

Analysis type : "UNI"

Data file : "data.txt"

Pedigree file : "pedi.txt"

Parameter file : "wombat.par"

 

 

No. of traits = 1

nrec mean sdev min. max.

1 "milk_yld" 42 16.2095 4.45283 8.00000 24.0000

 

Covariables

1 "dim(4,LEG)" 42 138.381 94.9724 4.00000 310.000

 

Fixed effects

1 "milk_yld" nlev

1 "htd" 10

 

Random effects nlev

1 "animal" 8 NRM

2 "pe_ani" 5 IDE

======== end of file ============================13-03-2014==========05:05====

 

SumPedigree.out

 

======= Version 19-05-2012 ======================================= **KM** ====

 

Program WOMBAT : Summary of Pedigree Information

==============================================================================

 

Fixed Regression Model from Mrode, 2nd Edition, Example 7.1

 

Analysis type : "UNI"

Data file : "data.txt"

Pedigree file : "pedi.txt"

Parameter file : "wombat.par"

 

 

No. of animal IDs in data file = = 5

No. of animal IDs in total = = 8

random effect no. = 1 NRM

no. of elements in NRM/GIN inverse 23

log determinant = -3.4657359027997265

random effect no. = 2 IDE

no. of elements in NRM/GIN inverse 0

log determinant = 0.

======== end of file ============================13-03-2014==========05:05====

 

FixSolutions.out

 

======= Version 19-05-2012 ======================================= **KM** ======================================

 

Program WOMBAT : GLS solutions for fixed effects

================================================================================================================

 

Fixed Regression Model from Mrode, 2nd Edition, Example 7.1

 

 

Covariables for trait no. 1 "milk_yld"

Covariable Reg.coeff Solution

1 dim(4,LEG) 1 -0.522879

1 dim(4,LEG) 2 -0.124463

1 dim(4,LEG) 3 0.535447

1 dim(4,LEG) 4 -0.419495

 

Fixed effects for trait no. 1 "milk_yld"

Effect Orig.code Level Solution SolSum=0 No.recs Eff.Mean

1 htd 1 1 6.30034 5.68808 3 20.733

1 htd 2 2 3.31709 2.70483 3 19.867

1 htd 3 3 4.02519 3.41293 3 21.000

1 htd 4 4 3.60258 2.99032 4 20.700

1 htd 5 5 1.70333 1.09107 4 18.850

1 htd 6 6 -1.48866 -2.10092 5 14.480

1 htd 7 7 -1.36809 -1.98035 5 14.860

1 htd 8 8 -1.28826 -1.90052 5 14.980

1 htd 9 9 -4.00293 -4.61519 5 12.160

1 htd 10 10 -4.67798 -5.29025 5 11.080

1 htd 0.612261

** marks effects which have been set to zero for the analysis

======== end of file ============================13-03-2014==========05:05======================================

 

RnSoln_animal.dat 확인

 

Run N Original ID Tr Solution Inbr %

1 1 1 -0.329992 0.000

2 2 1 -0.160408 0.000

3 3 1 0.490400 0.000

4 4 1 0.432813E-02 0.000

5 5 1 -0.244940 0.000

6 6 1 -0.836741 0.000

7 7 1 1.14770 0.000

8 8 1 0.378609 12.500

 

RnSoln_pe_ani.dat

 

Run N Original ID Tr Solution

1 4 1 -0.615620

2 5 1 -0.415134

3 6 1 -1.68533

4 7 1 2.80889

5 8 1 -0.928011E-01

 

 

관련 파일

 

09_Fixed_Reg.zip