R package:blupADC-Feature 8
Table of Contents
Oveview
👻In the previous section, we have given detailed description about the interface with DMU by run_DMU function. In this chapter, we will introduce the usage of run_BLUPF90 function.
Note: the usage of run_BLUPF90 and run_DMU is similar. Thus, we recommend to have a look at the usage of run_DMU function.
👉 Note: Package blupADC has encapsulated the basic module of BLUPF90 (renumf90, remlf90, airemlf90, and blupf90), if you want to use more modules, please download from websit (BLUPF90 download website).
For commercial use of BLUPF90, user must contact the author of BLUPF90 !!!
👉 Note: Package blupADC now supports object-oriented programming in running BLUPF90, which should be more easier in analysis, see more details!
Example
Single trait model - Pedigree BLUP
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter","Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter")), #random effect name
covariate_effect_name=NULL, #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="pedigree.txt", #name of relationship file
output_result_path=getwd() # output path
)
Single trait model - GBLUP
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter","Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter")), #random effect name
covariate_effect_name=NULL, #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
analysis_model="GBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="blupf90_genumeric", #name of relationship file
output_result_path=getwd() # output path
)
Single trait model - Single-step BLUP
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter","Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter")), #random effect name
covariate_effect_name=NULL, #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
analysis_model="SSBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name=c("pedigree.txt","blupf90_genumeric"), #name of relationship file
output_result_path=getwd() # output path
)
Similar to run_DMU function, through modifying the two parameters: analysis_model and relationship_name , we can perform Pedigree-BLUP, GBLUP, and SSBLUP analysis(PS: blupf90_genumeric is generated through cal_kinship function，see more details about cal_kinship function).
Multiple traits model - Pedigree BLUP
The following code is about the usage of multiple traits model through BLUPF90:
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter","Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1"),c("Trait2")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season"),c("Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter"),c("Id")), #random effect name
covariate_effect_name=list(NULL,"Age"), #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name=c("pedigree.txt"), #name of relationship file
output_result_path=getwd() # output path
)
Single trait - pedigree BLUP model(with user-provided prior)
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter",
"Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter")), #random effect name
covariate_effect_name=NULL, #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
provided_BLUPF90_prior_file_path=data_path, #path of user-provided prior file
provided_BLUPF90_prior_file_name="BLUPF90_Prior", #name of user-provided prior file
provided_BLUPF90_prior_effect_name=c("Id","Litter","Residual"),
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="pedigree.txt", #name of relationship file
output_result_path=getwd() # output path
)
Single trait - pedigree BLUP model( with maternal effect)
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Herd","B_month","D_age","Litter","Sex","HY","ID","DAM","L_Dam",
"W_birth","W_2mth","W_4mth","G_0_2","G_0_4","G_2_4"), # colnames of phenotype
target_trait_name=list(c("W_birth")), #trait name
fixed_effect_name=list(c("B_month","D_age","Litter","Sex","HY")), #fixed effect name
random_effect_name=list(c("ID","L_Dam")), #random effect name
maternal_effect_option=c("mat"),
genetic_effect_name="ID", #genetic effect name
covariate_effect_name=NULL, #covariate effect name
phe_path=data_path, #path of phenotype file
phe_name="maternal_data", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="maternal_pedigree", #name of relationship file
output_result_path=getwd() # output path
)
Single trait - pedigree BLUP model( with permanent effect)
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("id","year_grp","breed","time","t_dato",
"age","L1","L2","L3","gh"), # colnames of phenotype
target_trait_name=list(c("gh")), #trait name
fixed_effect_name=list(c("year_grp","breed","time")), #fixed effect name
random_effect_name=list(c("id","t_dato")), #random effect name
covariate_effect_name=list(c("age")), #covariate effect name
genetic_effect_name="id", #genetic effect name
included_permanent_effect=list(c(TRUE)), #whether include permant effect
phe_path=data_path, #path of phenotype file
phe_name="rr_data", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="rr_pedigree", #name of relationship file
output_result_path=getwd() # output path
)
Single trait - pedigree BLUP model( with random regression effect)
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("id","year_grp","breed","time","t_dato",
"age","L1","L2","L3","gh"), # colnames of phenotype
target_trait_name=list(c("gh")), #trait name
fixed_effect_name=list(c("year_grp","breed","time")), #fixed effect name
random_effect_name=list(c("id","t_dato")), #random effect name
covariate_effect_name=list(c("age")), #covariate effect name
genetic_effect_name="id", #genetic effect name
random_regression_effect_name=list(c("L1&id&pe_effect","L2&id&pe_effect")), #random regression effect name
phe_path=data_path, #path of phenotype file
phe_name="rr_data", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="rr_pedigree", #name of relationship file
output_result_path=getwd() # output path
)
Single trait model - BLUP (gibbs sampling)
library(blupADC)
data_path=system.file("extdata", package = "blupSUP") # path of provided files
run_BLUPF90(
phe_col_names=c("Id","Mean","Sex","Herd_Year_Season","Litter","Trait1","Trait2","Age"), # colnames of phenotype
target_trait_name=list(c("Trait1")), #trait name
fixed_effect_name=list(c("Sex","Herd_Year_Season")), #fixed effect name
random_effect_name=list(c("Id","Litter")), #random effect name
covariate_effect_name=NULL, #covariate effect name
genetic_effect_name="Id", #genetic effect name
phe_path=data_path, #path of phenotype file
phe_name="phenotype.txt", #name of phenotype file
analysis_model="PBLUP_A", #model of genetic evaluation
relationship_path=data_path, #path of relationship file
relationship_name="pedigree.txt", #name of relationship file
BLUPF90_algorithm="Gibbs",
output_result_path=getwd() # output path
)
Parameter
Many parameters in run_BLUPF90 are the same as in run_DMU function(see more details).
💫Thus, we will introduce specific parameters in run_BLUPF90 function.
1：BLUPF90_algorithm
Algorithm of estimating variance components,
characterclass. Default is “EM_REML”."AI_REML""EM_REML""BLUP": no need to estimate variance components, solve mixed linear model directly.
2：provided_blupf90_prior_file_path
File path of user-provided prior file, character class.
Note: The format of BLUPF90 prior file is different to the format of DMU prior file. In the next section, i will give a detailed introduction.
- 3：provided_blupf90_prior_file_name
File name of user-provided prior file, character class.
- 4：provided_blupf90_prior_effect_name
Effect name of user-provided prior file, character class.
- 5：BLUPf90_software_path
Path of software BLUPF90 , character class.
Quanshun Mei
Postdoctoral researcher
My research interests include applying genomic selection and machine learning in animal breeding.