BLUP

BLUP PCA

---

It is a novel method for calculating eigenvector for M SNPs in a sample with N individuals (given M>>N).
The method is a two-step procedure.

In step 1, the PCA is implemented on a A (N X N) matrix, and the individual eigenvector over the first K space can be calculated.
In step 2, the BLUP is used to estimate the effects for M SNPs for each SNP eigenvector.

Subcommand: bluppca

Available Options

--grm-bin, --grm or --grm-gz
Specify the genetic relationship matrix (GRM). The GRM is the A matrix mentioned in the algorithm, which quantifies the genetic relatedness between individuals over M SNPs. See GCTA document for details.

--file or --bfile
Specify the PLINK format data files. See PLINK document for details.

--pheno
Specify the phenotype file, i.e. the individual eigenvectors. The phenotype file should consist of at least 3 columns. The fist column is family IDs. The second column is individual IDs. Starting from the third column, each column records the phenotypes of all the individuals of a trait.

--out
Specify the output root filename. The result is outputted to a .blup file.

Example
gear bluppca --grm-bin test --bfile test --pheno test.eigenvec --out test
gear bluppca --grm test --bfile test --pheno test.eigenvec --out test