data format

genotype

279 0 0 1 1 1
280 1 0 0 2 2
281 1 0 1 0 1

phenotype

5 1.869
6 -1.140
7 1.253

pedigree

1 0 0
2 0 0
3 1 2
  • Founders are denoted as 0 in pedigree.
  • Animal IDs can be any string, don't have to be numbers.
  • Headers(marker IDs) for genotypes can be included.

input parameters

input=InputParameters()
fieldnames(input)

#13-element Array{Symbol,1}:
# :seed            
# :method          
# :chainLength     
# :outFreq         
# :probFixed       
# :varGenotypic    
# :varResidual     
# :estimateVariance
# :estimatePi      
# :estimateScale   
# :dfEffectVar     
# :nuRes           
# :nuGen

methods

Methods in the SSBR package:

  1. BayesC0_constantvariance
  2. BayesC0
  3. BayesB
  4. BayesC

Coming..

ssGBLUP exact ssGBLUP APY ssGBLUP

SSBR

SSBR is a tool for single step Bayesian regression analyses.

Quick-start

using JWAS: Datasets,SSBR,misc

#data files from QTLDatasets package
pedfile    = Datasets.dataset("testSSBR","ped.txt")
genofile   = Datasets.dataset("testSSBR","genotype.txt")
phenofile  = Datasets.dataset("testSSBR","phenotype.txt")
fixedfile  = Datasets.dataset("testSSBR","fixed.txt")
Validation = Datasets.dataset("testSSBR","validation.txt")

#set up input parameters
input=InputParameters()
input.method       = "BayesC"
input.varGenotypic = 4.48
input.varResidual  = 6.72
input.probFixed    = 0.99
input.outFreq      = 10000


MCMCinfo(input)
#MCMC Information:
#seed                        314
#chainLength               50000
#method                   BayesC
#outFreq                    1000
#probFixed                 0.990
#varGenotypic              4.480
#varResidual               6.720
#estimateVariance           true
#estimatePi                false
#estimateScale             false
#dfEffectVar               4.000
#nuRes                     4.000
#nuGen                     4.000
#centering                 false


#run it
out=runSSBR(input,pedigree=pedfile,genotype=genofile,phenotype=phenofile,fixedfile=fixedfile);

#check accuracy
using DataFrames
df = readtable(Validation, eltypes =[UTF8String, Float64], separator = ' ',header=false,names=[:ID,:EBV]);
comp=join(out,df,on=:ID);
cor(comp[:EBV],comp[:EBV_1])

More


In [ ]: