Eagle-package {Eagle}R Documentation

Eagle for Genome-wide Association Mapping


An implementation of multiple-locus association mapping on a genome-wide scale. 'Eagle' can handle inbred and outbred study populations, populations of arbitrary unknown complexity, and data larger than the memory capacity of the computer. Since 'Eagle' is based on linear mixed models, it is best suited to the analysis of data on continuous traits. However, it can tolerate non-normal data. 'Eagle' reports, as its findings, the best set of snp in strongest association with a trait. For users unfamiliar with R, to perform an analysis, run 'OpenGUI()'. This opens a web browser to the menu-driven user interface for the input of data, and for performing genome-wide analysis.


Motivation: Data from genome-wide association studies are analyzed, commonly, with single-locus models. That is, analyzes are performed on a locus-by-locus basis. Multiple-locus approaches that model the association between a trait and multiple loci simultaneously are more powerful. However, these methods do not scale well with study size and many of the packages that implement these methods are not easy to use. Eagle was specifically designed to make genome-wide association mapping with multiple-locus models simple and practical.


  1. Individuals are diploid but they can be inbred or outbred.

  2. The marker and phenotype data are in separate files.

  3. Marker loci are snps. Dominant and multi-allelic loci will need to be converted into biallelic (snp-like) loci.

  4. The trait is continuous and normally distributed. Eagle can handle non-normally distributed trait data but there may be a loss of power to detect marker-trait associations.

Important Functions:

  1. ReadMarker for reading in the snp data.

  2. ReadPheno for reading in the phenotypic data (traits and features/covariates)

  3. ReadMap for reading in the marker map.

  4. FPR4AM for calculating the value of the lambda parameter to be used by AM that will give a desired false positive rate for detecting SNP-trait associations.

  5. AM for performing association mapping on the data.

  6. OpenGUI which opens the GUI.

Output: The key output from AM is a list of snp. Each snp identifies a separate genomic region of interest, housing genes that are affecting the trait. Additional summary information such as the size of the snp effects, their statistical significance, and how much phenotypic variation they explain can be obtained by running SummaryAM.

Where to get help: A variety of different help options are available.


Andrew W. George (Data61, CSIRO) with a lot of support from Joshua Bowden (IM&T, CSIRO)

Maintainer: Andrew W. George <geo047@gmail.com>

[Package Eagle version 2.4.5 Index]