Error in solve.default(ww)

[Chase556677]

Hi Xiaolei,

When I use this command for the FarmCPU GWAS(showed below)

imMVP <- MVP(phe=phenotype,geno=genotype,map=map,K=Kinship,CV.FarmCPU=Covariates,nPC.FarmCPU=3,perc=1,priority="speed",ncpus=8,vc.method="EMMA",maxLoop=10,method.bin="FaST-LMM",permutation.threshold=TRUE,permutation.rep=100,threshold=0.05,method="FarmCPU")


#--------------------------------------Welcome to MVP--------------------------------------# 
#    A Memory-efficient, Visualization-enhanced, and Parallel-accelerated Tool For GWAS    # 
# Version: 1.0.1                                                                           # 
# Authors: Lilin Yin, Haohao Zhang, Zhiwu Zhang, Xinyun Li, Shuhong Zhao, and Xiaolei Liu  # 
# Contact: xiaoleiliu@mail.hzau.edu.cn                                                     # 
#------------------------------------------------------------------------------------------# 
[1] "Input data has 378 individuals, 258873 markers"
[1] "Principal Component Analysis Start..."
 means for first 10 snps:
 [1] 0 0 0 0 0 0 0 0 0 0
[1] "GWAS Start..."
[1] "FarmCPU Start ..."
[1] "Current loop: 1 out of maximum of 10"
[1] "seqQTN"
NULL
[1] "scanning..."
[1] "number of covariates in current loop is:"
[1] 8
Error in solve.default(ww) : 
  Lapack routine dgesv: system is exactly singular: U[5,5] = 0

I got an error about Error in solve.default(ww).The kinship and PCA were calculated by MVP.
Please help,Thanks.

[YinLiLin]

If the covariates are PCA, please shut down the parameter "nPC.FarmCPU", they are the same. So you could run as fallowing:

imMVP <- MVP(phe=phenotype,geno=genotype,map=map,K=Kinship,CV.FarmCPU=Covariates,
            perc=1,priority="speed",ncpus=8,vc.method="EMMA",maxLoop=10,method.bin=
           "FaSTLMM",permutation.threshold=TRUE,permutation.rep=100,threshold=0.05,
            method="FarmCPU")

or

imMVP <- MVP(phe=phenotype,geno=genotype,map=map,K=Kinship,nPC.FarmCPU=3,
            perc=1,priority="speed",ncpus=8,vc.method="EMMA",maxLoop=10,method.bin=
           "FaST-LMM",permutation.threshold=TRUE,permutation.rep=100,threshold=0.05,
            method="FarmCPU")

If the covariates are not PCA, please check your covariates, maybe the determinant of the matrix is too small to be inversed.

[hyacz]

Closed because there is no activity for too long. if you have any related new messages, feel free to open it.