# Linear Models for the Prediction of Animal Breeding Values, 3rd Edition.
# Raphael Mrode
# Example 11.2 p183
간단한 설명은 다음 포스팅을 참고한다.
Mixed Linear Model for Computing SNP Effects(SNP-BLUP Model), unweighted analysis
# Linear Models for the Prediction of Animal Breeding Values, 3rd Edition. # Raphael Mrode # Example 11.2 p183 SNP effect를 임의 효과로 다루어 SNP effect를 추정한다. 각 SNP effect의..
bhpark.tistory.com
Data
13 0 0 1 558 9 0.00179211 1.3571429 -0.3571429 0.2857143 0.2857143 -0.2857143 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857
14 0 0 1 722 13.4 0.00138504 0.3571429 -0.3571429 -0.7142857 -0.7142857 -0.2857143 0.7857143 -0.1428571 0.07142857 -0.1428571 -0.7857143
15 13 4 1 300 12.7 0.00333333 0.3571429 0.6428571 1.2857143 0.2857143 0.7142857 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857
16 15 2 1 73 15.4 0.01369863 -0.6428571 -0.3571429 1.2857143 0.2857143 -0.2857143 -0.2142857 -0.1428571 0.07142857 0.8571429 0.2142857
17 15 5 1 52 5.9 0.01923077 -0.6428571 0.6428571 0.2857143 1.2857143 -0.2857143 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857
18 14 6 1 87 7.7 0.01149425 0.3571429 0.6428571 -0.7142857 0.2857143 -0.2857143 0.7857143 -0.1428571 0.07142857 0.8571429 0.2142857
19 14 9 1 64 10.2 0.01562500 -0.6428571 -0.3571429 0.2857143 0.2857143 -0.2857143 0.7857143 -0.1428571 0.07142857 0.8571429 -0.7857143
20 14 9 1 103 4.8 0.00970874 -0.6428571 0.6428571 0.2857143 -0.7142857 -0.2857143 -0.2142857 -0.1428571 0.07142857 0.8571429 -0.7857143
1 ~ 3 : animal, sire, dam
4 : general mean
5 : EDC(using weight)
6 : Fat DYD
7 : EDC 역수
8 - 17 : SNP1 ~ SNP10의 coding하고 평균을 0으로 scaling한 값
(7 - 17 컬럼은 원래의 자료에서 계산을 하여 입력하여야 한다.)
* 계산 방법은 위 포스팅을 참고
Renumf90 Parameter File
# Parameter file for program renf90; it is translated to parameter
# file for BLUPF90 family programs.
DATAFILE
snp_data2.txt
TRAITS
6
FIELDS_PASSED TO OUTPUT
WEIGHT(S)
RESIDUAL_VARIANCE
245
EFFECT
4 cross alpha
EFFECT
4 cross alpha
RANDOM
diagonal
RANDOM_REGRESSION
data
RR_POSITION
8 9 10 11 12 13 14 15 16 17
(CO)VARIANCES
9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96
OPTION solv_method FSPAK
첫째
EFFECT
4 cross alpha
general mean의 추정을 위하여 넣은 effect
두번째
EFFECT
4 cross alpha
SNP 효과들을 회귀 변수로 다루기 위한 effect. 공변이를 임의 효과로 다루므로 일종의 random regression model이다. 이 효과에 대해서 nested 하는데 결국 모두 1이므로 전체에 대해서 회귀 계수(SNP effect)를 추정하게 된다.
RANDOM
diagonal
대각 성분에만 분산을 더하여 효과 추정
RANDOM_REGRESSION
data
공변이들이 자료에 있다는 의미
RR_POSITION
8 9 10 11 12 13 14 15 16 17
자료에 있는 공변이들의 위치
(CO)VARIANCES
9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96 0.00
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.96
10개 random SNP effect의 분산 행렬. 여기서 9.96은 상가적 유전분산(35.242)을 각 SNP의 2p(1-p)의 합(3.5382)으로 나누어준 값이다.
Renumf90 실행 화면
renumf90 실행 로드
RENUMF90 version 1.145
renumf90_mlm_snp_uw.par
datafile:snp_data2.txt
traits: 6
R
245.0
Processing effect 1 of type cross
item_kind=alpha
Processing effect 2 of type cross
item_kind=alpha
Reading (CO)VARIANCES: 10 x 10
Maximum size of character fields: 20
Maximum size of record (max_string_readline): 800
Maximum number of fields for input file (max_field_readline): 100
Pedigree search method (ped_search): convention
Order of pedigree animals (animal_order): default
Order of UPG (upg_order): default
Missing observation code (missing): 0
hash tables for effects set up
first 3 lines of the data file (up to 70 characters)
13 0 0 1 558 9 0.00179211 1.3571429 -0.3571429 0.2857143 0.2857143 -0.
14 0 0 1 722 13.4 0.00138504 0.3571429 -0.3571429 -0.7142857 -0.714285
15 13 4 1 300 12.7 0.00333333 0.3571429 0.6428571 1.2857143 0.2857143
read 8 records
table with 1 elements sorted
added count
Effect group 1 of column 1 with 1 levels
table expanded from 10000 to 10000 records
table with 1 elements sorted
added count
Effect group 2 of column 1 with 1 levels
table expanded from 10000 to 10000 records
wrote statistics in file "renf90.tables"
Basic statistics for input data (missing value code is '0')
Pos Min Max Mean SD N
6 4.8000 15.400 9.8875 3.7434 8
random effect 2
type:diag
Wrote parameter file "renf90.par"
Wrote renumbered data "renf90.dat" 8 records
renumf90으로 생성된 파일
renf90.dat
9 1.3571429 -0.3571429 0.2857143 0.2857143 -0.2857143 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857 1 1
13.4 0.3571429 -0.3571429 -0.7142857 -0.7142857 -0.2857143 0.7857143 -0.1428571 0.07142857 -0.1428571 -0.7857143 1 1
12.7 0.3571429 0.6428571 1.2857143 0.2857143 0.7142857 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857 1 1
15.4 -0.6428571 -0.3571429 1.2857143 0.2857143 -0.2857143 -0.2142857 -0.1428571 0.07142857 0.8571429 0.2142857 1 1
5.9 -0.6428571 0.6428571 0.2857143 1.2857143 -0.2857143 -1.2142857 -0.1428571 0.07142857 -0.1428571 1.2142857 1 1
7.7 0.3571429 0.6428571 -0.7142857 0.2857143 -0.2857143 0.7857143 -0.1428571 0.07142857 0.8571429 0.2142857 1 1
10.2 -0.6428571 -0.3571429 0.2857143 0.2857143 -0.2857143 0.7857143 -0.1428571 0.07142857 0.8571429 -0.7857143 1 1
4.8 -0.6428571 0.6428571 0.2857143 -0.7142857 -0.2857143 -0.2142857 -0.1428571 0.07142857 0.8571429 -0.7857143 1 1
renf90.par
# BLUPF90 parameter file created by RENUMF90
DATAFILE
renf90.dat
NUMBER_OF_TRAITS
1
NUMBER_OF_EFFECTS
11
OBSERVATION(S)
1
WEIGHT(S)
EFFECTS: POSITIONS_IN_DATAFILE NUMBER_OF_LEVELS TYPE_OF_EFFECT[EFFECT NESTED]
12 1 cross
2 1 cov 13
3 1 cov 13
4 1 cov 13
5 1 cov 13
6 1 cov 13
7 1 cov 13
8 1 cov 13
9 1 cov 13
10 1 cov 13
11 1 cov 13
RANDOM_RESIDUAL VALUES
245.00
RANDOM_GROUP
2 3 4 5 6 7 8 9 10 11
RANDOM_TYPE
diagonal
FILE
(CO)VARIANCES
9.9600 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000 0.0000 0.0000
0.0000 9.9600 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000 0.0000 0.0000
0.0000 0.0000 9.9600 0.0000 0.0000 0.0000 0.0000
0.0000 0.0000 0.0000
0.0000 0.0000 0.0000 9.9600 0.0000 0.0000 0.0000
0.0000 0.0000 0.0000
0.0000 0.0000 0.0000 0.0000 9.9600 0.0000 0.0000
0.0000 0.0000 0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 9.9600 0.0000
0.0000 0.0000 0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 9.9600
0.0000 0.0000 0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
9.9600 0.0000 0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000 9.9600 0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000 0.0000 9.9600
OPTION solv_method FSPAK
위에서 생성된 renf90.dat와 renf90.par를 이용하여 blupf90 실행
blupf90 실행 화면
blupf90 실행 로그
renf90.par
BLUPF90 ver. 1.68
Parameter file: renf90.par
Data file: renf90.dat
Number of Traits 1
Number of Effects 11
Position of Observations 1
Position of Weight (1) 0
Value of Missing Trait/Observation 0
EFFECTS
# type position (2) levels [positions for nested]
1 cross-classified 12 1
2 covariable 2 1 13
3 covariable 3 1 13
4 covariable 4 1 13
5 covariable 5 1 13
6 covariable 6 1 13
7 covariable 7 1 13
8 covariable 8 1 13
9 covariable 9 1 13
10 covariable 10 1 13
11 covariable 11 1 13
Residual (co)variance Matrix
245.00
correlated random effects 2 3 4 5 6 7 8 9 10 11
Type of Random Effect: diagonal
trait effect (CO)VARIANCES
1 2 9.960 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
1 3 0.000 9.960 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
1 4 0.000 0.000 9.960 0.000 0.000 0.000 0.000 0.000 0.000 0.000
1 5 0.000 0.000 0.000 9.960 0.000 0.000 0.000 0.000 0.000 0.000
1 6 0.000 0.000 0.000 0.000 9.960 0.000 0.000 0.000 0.000 0.000
1 7 0.000 0.000 0.000 0.000 0.000 9.960 0.000 0.000 0.000 0.000
1 8 0.000 0.000 0.000 0.000 0.000 0.000 9.960 0.000 0.000 0.000
1 9 0.000 0.000 0.000 0.000 0.000 0.000 0.000 9.960 0.000 0.000
1 10 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 9.960 0.000
1 11 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 9.960
REMARKS
(1) Weight position 0 means no weights utilized
(2) Effect positions of 0 for some effects and traits means that such
effects are missing for specified traits
* The limited number of OpenMP threads = 4
* solving method (default=PCG):FSPAK
Data record length = 13
# equations = 11
G
9.9600 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000
0.0000 9.9600 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000
0.0000 0.0000 9.9600 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000
0.0000 0.0000 0.0000 9.9600 0.0000 0.0000 0.0000 0.0000 0.0000
0.0000
0.0000 0.0000 0.0000 0.0000 9.9600 0.0000 0.0000 0.0000 0.0000
0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 9.9600 0.0000 0.0000 0.0000
0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 9.9600 0.0000 0.0000
0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 9.9600 0.0000
0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 9.9600
0.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
9.9600
read 8 records in 0.2343750 s, 66
nonzeroes
finished peds in 0.2343750 s, 66 nonzeroes
left hand side
0.0327 -0.0006 0.0047 0.0093 0.0052 -0.0052 -0.0070 -0.0047 0.0023 0.0117 0.0070
-0.0006 0.1162 -0.0021 -0.0042 -0.0016 0.0016 -0.0040 0.0001 -0.0000 -0.0063 0.0080
0.0047 -0.0021 0.1092 0.0013 0.0028 0.0013 -0.0051 -0.0007 0.0003 0.0017 0.0051
0.0093 -0.0042 0.0013 0.1194 0.0056 0.0026 -0.0142 -0.0013 0.0007 0.0033 0.0102
0.0052 -0.0016 0.0028 0.0056 0.1130 -0.0003 -0.0093 -0.0007 0.0004 -0.0002 0.0134
-0.0052 0.0016 0.0013 0.0026 -0.0003 0.1048 -0.0030 0.0007 -0.0004 -0.0039 0.0030
-0.0070 -0.0040 -0.0051 -0.0142 -0.0093 -0.0030 0.1264 0.0010 -0.0005 0.0057 -0.0219
-0.0047 0.0001 -0.0007 -0.0013 -0.0007 0.0007 0.0010 0.1011 -0.0003 -0.0017 -0.0010
0.0023 -0.0000 0.0003 0.0007 0.0004 -0.0004 -0.0005 -0.0003 0.1006 0.0008 0.0005
0.0117 -0.0063 0.0017 0.0033 -0.0002 -0.0039 0.0057 -0.0017 0.0008 0.1127 -0.0057
0.0070 0.0080 0.0051 0.0102 0.0134 0.0030 -0.0219 -0.0010 0.0005 -0.0057 0.1264
right hand side:
0.32 0.00 0.01 0.12 0.04 -0.04 -0.05 -0.05 0.02 0.11
0.07
solution:
9.94 0.09 -0.31 0.26 -0.08 0.11 0.14 -0.00 0.00 -0.06
-0.02
solutions stored in file: "solutions"
blupf90 실행 결과 : solutions
trait/effect level solution
1 1 1 9.94394207
1 2 1 0.08702092
1 3 1 -0.31079215
1 4 1 0.26246002
1 5 1 -0.08027711
1 6 1 0.11020813
1 7 1 0.13908022
1 8 1 -0.00000000
1 9 1 0.00000000
1 10 1 -0.06069044
1 11 1 -0.01580233
general mean effect(1개) 추정하고, 10개의 SNP effect를 추정