Getting Started with selection.index"
In selection.index: Analysis of Selection Index in Plant Breeding

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Introduction

The selection.index package provides an optimized suite of tools to calculate multi-trait phenotypic, genomic, and multistage selection indices. It simplifies plant and animal breeding workflows by enabling breeders to accurately rank genotypes based on user-defined economic weights and experimental designs.

Setup

First, load the package and the built-in phenotypic dataset. While the package supports any standard multi-environment phenotypic data, we will use the included seldata dataset (a built-in multi-environment trial dataset for genotypes) for this quick-start guide.

library(selection.index)

# Load the built-in phenotypic dataset
data("seldata")

# Inspect the structure of the dataset
head(seldata)

The Basics

To calculate a selection index, you must specify the traits of interest, assign economic weights, and compute the genotypic and phenotypic variance-covariance matrices.

# Define economic weights for the 7 traits of interest
weights <- c(10, 8, 6, 4, 2, 1, 1)

# Calculate genotypic and phenotypic variance-covariance matrices
# Traits: columns 3:9, Genotypes: column 2, Replication: column 1
gmat <- gen_varcov(data = seldata[, 3:9], genotypes = seldata[, 2], replication = seldata[, 1])
pmat <- phen_varcov(data = seldata[, 3:9], genotypes = seldata[, 2], replication = seldata[, 1])

Execution

With the matrices and weights defined, pass them into the primary selection function. We use lpsi() to compute the Combinatorial Linear Phenotypic Selection Index for all traits. Here, setting ncomb = 7 calculates the index considering all 7 traits simultaneously.

# Calculate the combinatorial selection index for all 7 traits
index_results <- lpsi(
  ncomb = 7,
  pmat = pmat,
  gmat = gmat,
  wmat = as.matrix(weights),
  wcol = 1
)

Results

Once the index is computed, review the genetic advance metrics and extract the final selection scores to identify the top-performing genotypes. index_results is a data frame containing the index coefficients (b columns) and various genetic metrics (GA, PRE, Delta_G, rHI).

# View the calculated index metrics for our 7-trait combination
head(index_results)

# Extract the final selection scores to rank the genotypes
scores <- predict_selection_score(
  index_results,
  data = seldata[, 3:9],
  genotypes = seldata[, 2]
)

# View the top ranked genotypes based on their selection scores
head(scores)