Manual

(1) Make score file.

java -jar calsc.jar [covariate & phenotype file] [output score file] [logistic/linear]
ex) java -jar calsc.jar covphe.txt score_file.txt logistic

(2) Run GWAS-GMDR (CPU version)

gwasgmdr --in [input file] --out [output file] --order [order] --cv [cv fold] --measure [measure] --top [result top] --score [score file]
ex) gwasgmdr --in sample.txt --out result.csv --order 2 --cv 10 --measure ba --top 1000 --score score_file.txt

If you don't define '--score' option, this gwasgmdr works as MDR not GMDR.
ex) gwasgmdr --in sample.txt --out result.csv --order 2 --cv 10 --measure ba --top 1000

(3) Run GWAS-GMDR (GPU version)

1) Make binarized data file
    GMDRconverter [input data file] [output binarized file]
    ex) GMDRconverter sample.mdr sample.bin

2) Run GWAS-GMDR (GPU version)
    gwasgmdr --in [input binarized file] --out [output result file] --thread [#threads] --block [#block] --gpu [#gpu] --order [order] --cv [cv fold] --measure [measure] --top [result top] --score [score file]
    ex) cuGwam --in sample.bin --out result.csv --thread 440 --block 150 --gpu 3 --order 2 --cv 10 --measure ba --top 1000 --score score_file.txt

[Parameter]

--in : input file path + name

--out : result file path + name

--order : combination order (1,2,3,4,.... or 1-3, 2-5) (default : 2)

--cv : cross validation fold (2,3,4,....) (default : 10)

--measure : measure for snp selection (default : ba)

ba -> Balanced Accuracy
accuracy -> Accuracy
homogeneity -> Homogeneity Test Stat
gamma -> Gamma Stat
tau-b -> tau_b
tau-c -> tau_c
somers_d1 -> somers_d1
somers_d2 -> somers_d2
nmi -> nmi
lr -> lr

--missing : missing method (default : available)

none -> None
mc -> Missing Category
complete -> Complete
available -> Available

--tie : tie (0: High 1: Low 2: Ignore) (default : 0)

--top : K=number of final significant snps for presentation (default : 10)

--shuffling : shuffling (0 : not shuffling 1: shuffling) (default : 0)

--score : score file path + name

Example Data

Covariates and phenotype file : covphe.txt

SNP size 10, Sample size 5000 : marker_10.txt

SNP size 100, Sample size 5000 : marker_100.txt

SNP size 1000, Sample size 5000 : marker_1k.txt