# R-Package: TSDFGS

### Training Set Determination for Genomic Selection

## Overview

This R package includes different methods for determining training set for genomic selection using genetic algorithms (Holland JH, 1975) or simple exchange algorithms. Three different criteria included, which are r-score (Ou JH et al., 2018), PEV-score (Akdemir D et al., 2015), and CD-score (Laloe D, 1993). Phenotype data for the candidate set is not necessary for all these methods.

Core functions are written in Rcpp with Eigen linear algebra library which makes functions perform great.

## Installation

This package had uploaded to CRAN. You may directly install and call functions in the R console.

install.packages("TSDFGS")

library(TSDFGS)

## Functions

r_score(x, x0)

A criterion for finding training set which derived from Pearson's correlation between GEBVs (genomic estimated breeding value) and phenotype value of a test set.

Argument:

x: A genetic matrix for the training set

x0: A genetic matrix for the test set

Return value:

A numeric score

pev_score(x, x0)

PEV-score (Akdemir D et al., 2015) is a criterion for finding a training set which derived from the covariance of prediction of the test set.