Multi-Purpose Core Subset Selection

Core Hunter is a tool to sample diverse, representative subsets from large germplasm collections, with minimum redundancy. Such so-called core collections have applications in plant breeding and genetic resource management in general. Core Hunter can construct cores based on genetic marker data, phenotypic traits or precomputed distance matrices, optimizing one of many provided evaluation measures depending on the precise purpose of the core (e.g. high diversity, representativeness, or allelic richness). In addition, multiple measures can be simultaneously optimized as part of a weighted index to bring the different perspectives closer together. The Core Hunter library is implemented in Java 8 as an open source project (see < http://www.corehunter.org>).


Latest release

CRAN version Coverage Status Build Status

Development snapshot

Coverage Status Build Status

Core Hunter is a tool to sample diverse, representative subsets from large germplasm collections, with minimum redundancy. Such so-called core collections have applications in plant breeding and genetic resource management in general. Core Hunter can construct cores based on genetic marker data, phenotypic traits or precomputed distance matrices, optimizing one of many provided evaluation measures depending on the precise purpose of the core (e.g. high diversity, representativeness, or allelic richness). In addition, multiple measures can be simultaneously optimized as part of a weighted index to bring the different perspectives closer together. The Core Hunter library is implemented in Java 8 as an open source project (see http://www.corehunter.org).

Version 3 has been recoded from scratch using the JAMES framework which provides the applied optimization algorithms. Requirements

A Java Runtime Environment (JRE) version 8 or later is required to run Core Hunter.

Getting started

The package corehunter can be installed from CRAN with

All provided functions are documented in the package, including many examples, for example try

> ?corehunter
> ?sampleCore
> ?genotypes
> ?phenotypes
> ?distances

For more information please visit http://www.corehunter.org.

Supported data types

Core Hunter 3 supports multiple types of genetic marker data, phenotypic traits and precomputed distance matrices. See http://www.corehunter.org/data for more details. Data can be loaded from files, data frames and matrices.

Evaluation measures

One of the main strengths of Core Hunter is that it can directly optimize a number of different evaluation measures. If desired, multiple measures can be simultaneously optimized as part of a weighted index. The measures included in Core Hunter 3 are listed below.

Distance based measures

  • Average entry-to-nearest-entry distance (diversity)
  • Average accession-to-nearest-entry distance (representativeness)
  • Average entry-to-entry distance (provided for historical reasons, not preferred)

Gower's distance is used to compute distances from phenotypic traits, and both the Modified Roger's as well as Cavalli-Sforza & Edwards distances are supported for genetic marker data. Alternatively, a precomputed distance matrix can be used.

Allelic richness

  • Shannon's index
  • Expected heterozygosity
  • Allele coverage

Available for genetic marker data only.

News

Core Hunter 3

Version 3.2.0 (08/09/2017)

  • New option to specify a set of always and/or never selected accessions.
  • Added support for step-based in addition to time-based stop conditions.
  • Using set.seed prior to executing Core Hunter, in combination with step-based stop conditions, now yields reproducible results.
  • Largely reduced memory footprint of genotype and distance data.
  • Default maximum time without improvement is now only applied in case no explicit stop conditions have been specified.
  • Fixed issue when loading phenotype data with a single trait.

Version 3.1.0 (27/01/2017)

  • More informative output when printing Core Hunter data objects (#19, #23).
  • It is now required to explicitly specify the format when reading genotype data. Moreover, a warning is raised in case it seems that the wrong format may have been selected, through an inspection of the data that was read (#20).
  • Fixed issues when reading files with single or double quoted values.
  • Improved test coverage.

Version 3.0.1 (04/10/2016)

  • The available Java version is now checked when installing or loading the package. Core Hunter requires Java 8 or higher.
  • Package title and description have been updated.

Version 3.0.0 (03/10/2016)

  • Initial release of Core Hunter 3 package

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("corehunter")

3.2.1 by Herman De Beukelaer, 8 days ago


Browse source code at https://github.com/cran/corehunter


Authors: Herman De Beukelaer [aut, cre], Guy Davenport [aut], Veerle Fack [ths]


Documentation:   PDF Manual  


MIT + file LICENSE license


Imports naturalsort, methods

Depends on rJava

Suggests testthat, mockr, StatMatch

System requirements: Java JRE 8 or higher


See at CRAN