全基因组选择中的R包

心血来潮，收集了全基因组选择中涉及的R包，有些是我经常使用的，有些是从CRAN的介绍中找出来的，这些包都和GS相关。

先立个Flag，我会将这些包逐一介绍相关使用说明及注意事项。

rrBLUP

rrBLUP
Software for genomic prediction with the RR-BLUP mixed model (Endelman 2011, [doi:10.3835/plantgenome2011.08.0024](https://doi.org/10.3835/plantgenome2011.08.0024)). One application is to estimate marker effects by ridge regression; alternatively, BLUPs can be calculated based on an additive relationship matrix or a Gaussian kernel.

sommer

sommer
Structural multivariate-univariate linear mixed model solver for multiple random effects allowing the specification of variance-covariance structures for random effects and allowing the fit of heterogeneous variance models (Covarrubias-Pazaran, 2016 [doi:10.1371/journal.pone.0156744](https://doi.org/10.1371/journal.pone.0156744); Maier et al., 2015 [doi:10.1016/j.ajhg.2014.12.006](https://doi.org/10.1016/j.ajhg.2014.12.006)). ML/REML estimates can be obtained using the Direct-Inversion Newton-Raphson, and Efficient Mixed Model Association algorithms. Designed for genomic prediction and genome wide association studies (GWAS), particularly focused in the p > n problem (more coefficients than observations) to include multiple relationship matrices or other covariance structures. Spatial models can be fitted using the two-dimensional spline functionality in sommer.

synbreed

synbreed
Data sets for the ‘synbreed’ package with three data sets from cattle, maize and mice to illustrate the functions in the ‘synbreed’ R package. All data sets are stored in the gpData format introduced in the ‘synbreed’ package. This research was funded by the German Federal Ministry of Education and Research (BMBF) within the AgroClustEr Synbreed - Synergistic plant and animal breeding (FKZ 0315528A).

cpgen

cpgen
Frequently used methods in genomic applications with emphasis on parallel computing (OpenMP). At its core, the package has a Gibbs Sampler that allows running univariate linear mixed models that have both, sparse and dense design matrices. The parallel sampling method in case of dense design matrices (e.g. Genotypes) allows running Ridge Regression or BayesA for a very large number of individuals. The Gibbs Sampler is capable of running Single Step Genomic Prediction models. In addition, the package offers parallelized functions for common tasks like genome-wide association studies and cross validation in a memory efficient way.

BGGE

BGGE

BGGE: Bayesian Genomic Linear Models Applied to GE Genome Selection
Application of genome prediction for a continuous variable, focused on genotype by environment (GE) genomic selection models (GS). It consists a group of functions that help to create regression kernels for some GE genomic models proposed by Jarquín et al. (2014) [doi:10.1007/s00122-013-2243-1](https://doi.org/10.1007/s00122-013-2243-1) and Lopez-Cruz et al. (2015) [doi:10.1534/g3.114.016097](https://doi.org/10.1534/g3.114.016097). Also, it computes genomic predictions based on Bayesian approaches. The prediction function uses an orthogonal transformation of the data and specific priors present by Cuevas et al. (2014) [doi:10.1534/g3.114.013094](https://doi.org/10.1534/g3.114.013094).

BGLR

BGLR
Application of genome prediction for a continuous variable, focused on genotype by environment (GE) genomic selection models (GS). It consists a group of functions that help to create regression kernels for some GE genomic models proposed by Jarquín et al. (2014) [doi:10.1007/s00122-013-2243-1](https://doi.org/10.1007/s00122-013-2243-1) and Lopez-Cruz et al. (2015) [doi:10.1534/g3.114.016097](https://doi.org/10.1534/g3.114.016097). Also, it computes genomic predictions based on Bayesian approaches. The prediction function uses an orthogonal transformation of the data and specific priors present by Cuevas et al. (2014) [doi:10.1534/g3.114.013094](https://doi.org/10.1534/g3.114.013094).

GSMX

GSMX
Estimating trait heritability and handling overfitting. This package includes a collection of functions for (1) estimating genetic variance-covariances and calculate trait heritability; and (2) handling overfitting by calculating the variance components and the heritability through cross validation.

Popvar

Popvar
The main attribute of ‘PopVar’ is the prediction of genetic variance in bi-parental populations, from which the package derives its name. ‘PopVar’ contains a set of functions that use phenotypic and genotypic data from a set of candidate parents to 1) predict the mean, genetic variance, and superior progeny value of all, or a defined set of pairwise bi-parental crosses, and 2) perform cross-validation to estimate genome-wide prediction accuracy of multiple statistical models. More details are available in Mohammadi, Tiede, and Smith (2015). Crop Sci. doi:10.2135/cropsci2015.01.0030. A dataset ‘think_barley.rda’ is included for reference and examples.

snnR

snnR

Solve the problem of over-parameterization in neural networks for genomic selection. Daniel Gianola, Hayrettin OkutEmail, Kent A Weigel and Guilherme JM Rosa (2011) [doi:10.1186/1471-2156-12-87](https://doi.org/10.1186/1471-2156-12-87).

xbreed

xbreed
xbreed: Genomic Simulation of Purebred and Crossbred Populations

Simulation of purebred and crossbred genomic data as well as pedigree and phenotypes are possible by this package. ‘xbreed’ can be used for the simulation of populations with flexible genome structures and trait genetic architectures. It can also be used to evaluate breeding schemes and generate genetic data to test statistical tools.

snpReady

snpReady
snpReady: Preparing Genotypic Datasets in Order to Run Genomic Analysis

Three functions to clean, summarize and prepare genomic datasets to Genome Selection and Genome Association analysis and to estimate population genetic parameters.