Abstract
The purpose of this chapter is to present recent advances in models for genomic-enabled prediction developed for ordinal categorical and count data. For both models we provide details of their corresponding derivation and then apply them to a real data set. The proposed models were derived using a Bayesian framework. Bayesian logistic ordinal regression (BLOR) and Bayesian negative binomial regression (BNBR) make use of the Pólya-Gamma distribution to produce an analytic Gibbs, a sampler with similar full conditional distributions of a model with Gaussian response and can be used for complex data sets as those that arise in the context of genomic selection where the sample size usually is smaller than the number of covariates (markers). We illustrate the proposed models using simulation and a real data set. Results indicate that our models for ordinal categorical and count data are a good alternative for analyzing ordinal and count data in the context of genomic-enabled prediction.
References
Albert JH, Chib S (1993) Bayesian analysis of binary and polychotomous response data. J Am Stat Assoc 88(422):669–679
Berridge DM, Crouchley R (2011) Multivariate generalized linear mixed models using R. CRC Press, Boca Raton
Bartlett MS (1947) The use of transformations. Biometrics 3(1):39–52
Brier GW (1950) Verification of forecasts expressed in terms of probability. Mon Weather Rev 78:1–3
Burgueño J, de los Campos GDL, Weigel K, Crossa J (2012) Genomic prediction of breeding values when modeling genotype × environment interaction using pedigree and dense molecular markers. Crop Sci 52:707–719
Casellas J, Caja G, Ferret A, Piedrafita J (2007) Analysis of litter size and days to lambing in the Ripollesa ewe. I. comparison of models with linear and threshold approaches . J Anim Sci 85:618–624
Cavanagh, C.R., Chao, S., Wang, S. et al. (2013). Genome-wide comparative diversity uncovers multiple targets of selection for improvement in hexaploid wheat landraces and cultivars. Proceedings of the National Academy of Sciences. 110(20):8057–8062
Crossa J, Pérez-Rodríguez P, de los Campos G, Mahuku G, Dreisigacker S, Magorokosho C (2011) Genomic selection and prediction in plant breeding. Journal of Crop Improvement 25(3):239–261
Czado C, Gneiting T, Held L (2009) Predictive model assessment for count data. Biometrics 65(4):1254–1261
de los Campos, G., and Perez-Rodriguez, P. (2013). BGLR: Bayesian generalized linear regression. R package version. http://R-Forge.R-project.org/projects/bglr/
de Maturana EL, Gianola D, Rosa GJM, Weigel KA (2009) Predictive ability of models for calving difficulty in US Holsteins. J Anim Breed Genet 126:179–188
Garthwaite PH, Kadane JB, O'Hagan A (2005) Statistical methods for eliciting probability distributions. J Am Stat Assoc 100(470):680–701
Gelfand AE, Smith AF (1990) Sampling-based approaches to calculating marginal densities. J Am Stat Assoc 85(410):398–409
Geyer CJ (1992) Practical Markov chain Monte Carlo. Stat Sci 7(4):473–483
Gianola D (1980) A method of sire evaluation for dichotomies. J of Anim Sci 51(6):1266–1271
Gianola D (1982) Theory and analysis of threshold characters. J Anim Sci 54(5):1079–1096
Gianola D, Foulley JL (1983) Sire evaluation for ordered categorical data with a threshold model. Genet Sel Evol 15(2):1–23
Gianola D (2013) Priors in whole-genome regression: the Bayesian alphabet returns. Genetics 194:573–596
González-Camacho JM, de los Campos G, Pérez-Rodríguez P, Gianola D, Cairns JE, Mahuku G, Crossa J (2012) Genome-enabled prediction of genetic values using radial basis function neural networks. Theor Appl Genet 125(4):759–771
González-Recio O, Forni S (2011) Genome-wide prediction of discrete traits using Bayesian regressions and machine learning. Genet Sel Evol 43:7
Hoerl AE, Kennard RW (1970) Ridge regression: biased estimation for nonorthogonal problems. Technometrics 12(1):55–67
Kizilkaya K, Tait RG, Garrick DJ, Fernando RL, Reecy JM (2011) Whole genome analysis of infectious bovine keratoconjunctivitis in Angus cattle using Bayesian threshold models. BMC Proc 5:S22
Kizilkaya K, Fernando RL, Garrick DJ (2014) Reduction in accuracy of genomic prediction for ordered categorical data compared to continuous observations. Genet Sel Evol 46(1):37. doi: 10.1186/1297-9686-46-37
Link WA, Eaton MJ (2012) On thinning of chains in MCMC. Methods Ecol Evol 3(1):112–115
MacEachern SN, Berliner LM (1994) Subsampling the Gibbs sampler. Am Stat 48(3):188–190
McCulloch CE, Searle SR (2001) Generalized, linear, and mixed models (1st ed.). Chichester: Wiley. ISBN 0-471-19364-X.
Montesinos-López OA, Montesinos-López A, Pérez-Rodríguez P, de los Campos G, Eskridge KM, Crossa J (2015a) Threshold models for genome-enabled prediction of ordinal categorical traits in plant breeding. G3: Genes| Genomes| Genetics 5(1):291–300
Montesinos-López OA, Montesinos-López A, Crossa J, Burgueño J, Eskridge K (2015b) Genomic-enabled prediction of ordinal data with Bayesian logistic ordinal regression. G3: Genes|Genomes|Genetics 5(10):2113–2126. http://doi.org/10.1534/g3.115.021154
Montesinos-López OA, Montesinos-López A, Pérez-Rodríguez P, Eskridge K, He X, Juliana P, Crossa J (2015c) Genomic prediction models for count data. J Agric Biol Environ Stat 20(2):533–554
Montesinos-López A, Montesinos-López OA, Crossa J, Burgueño J, Eskridge K, Falconi-Castillo E, He X, Singh P, Cichy K (2016) Genomic Bayesian prediction model for count data with genotype × environment interaction. G3: Genes|Genomes|Genetics 6(5):1165–1177
Nelder JA, Wedderburn RWM (1972) Generalized linear models. J R Stat Soc A 135:370–384. doi:10.2307/2344614
O’Hara RB, Kotze DJ (2010) Do not log-transform count data. Methods Ecol Evol 1(2):118–122
Park T, van Dyk DA (2009) Partially collapsed Gibbs samplers: illustrations and applications. J Comput Graph Stat 18(2):283–305
Polson NG, Scott JG, Windle J (2013) Bayesian inference for logistic models using Pólya–gamma latent variables. J Am Stat Assoc 108:1339–1349
Quenouille MH (1949) A relation between the logarithmic, Poisson, and negative binomial series. Biometrics 5:162–164
Ramirez-Valverde R, Misztal I, Bertrand J, K. (2001) Comparison of threshold vs linear and animal vs sire models for predicting direct and maternal genetic effects on calving difficulty in beef cattle. J Anim Sci 79:333–338
R Core Team (2015) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900051–07-0, URL http://www.R-project.org/
Schurink A, Wolc A, Ducro B, Frankena K, Garrick D, Dekkers J, van Arendonk J (2012) Genome-wide association study of insect bite hypersensitivity in two horse populations in the Netherlands. Genet Sel Evol 44(1):31
Scott J, Pillow JW (2013) Fully Bayesian inference for neural models with negative-binomial spiking. In: Pereira F, Burges CJC, Bottou L, Weinberger KQ (eds) Advances in neural information processing systems 25. Cornell University, New York, pp 1898–1906
Sorensen DA, Andersen S, Gianola D, Korsgaard I (1995) Bayesian inference in threshold models using Gibbs sampling. Genet Sel Evol 27(3):229–249
Stroup WW (2012) Generalized linear mixed models: modern concepts, methods and applications. CRC Press, Boca Raton
Stroup WW (2015) Rethinking the analysis of non-normal data in plant and soil science. Agron J 107(2):811–827
Teerapabolarn K, Jaioun K (2014) An improved Poisson approximation for the negative binomial distribution. Appl Math Sci 8(89):4441–4445
VanRaden PM (2008) Efficient methods to compute genomic predictions. J Dairy Sci 91(11):4414–4423
Vazquez AI, Weigel KA, Gianola D, Bates DM, Perez-Cabal MA et al (2009) Poisson versus threshold models for genetic analysis of clinical mastitis in US Holsteins. J Dairy Sci 92:5239–5247
Varona L, Misztal I, Bertrand J, K. (1999) Threshold-linear versus linear-linear analysis of birth weight and calving ease using an animal model. Ii. Comparison of models. J Anim Sci 77:2003–2007
Villanueva B, Fernandez J, Garcia-Cortes LA, Varona L, Daetwyler HD, Toro MA (2011) Accuracy of genome-wide evaluation for disease resistance in aquaculture breeding programs. J Anim Sci 89:3433–3442
Wang CL, Ding XD, Wang JY, Liu JF, Fu WX, Zhang Z, Jin ZJ, Zhang Q (2013) Bayesian methods for estimating GEBVs of threshold traits. Heredity 110(3):213–219
Wecker WE (1989) Assessing the accuracy of time series model forecasts of count observations. J Bus Econ Stat 7(4):418–419
Wright S (1934) An analysis of variability in number of digits in an inbred strain of guinea pigs. Genetics 19:506–536
Yang W, Tempelman RJ (2012) A Bayesian antedependence model for whole genome prediction. Genetics 190(4):1491–1501
Zucknick, M., and Richardson, S. (2014). MCMC algorithms for Bayesian variable selection in the logistic regression model for large-scale genomic applications. Technical Report. http://arxiv.org/abs/1402.2713.
Zhou M, Li L, Dunson D, Carin L (2012) Lognormal and gamma mixed negative binomial regression. In machine learning: proceedings of the international conference on machine learning. vol. 2012. p 1343. NIH Public Access.
Acknowledgments
We would like to thank all researchers in CIMMYT’s Global Maize Program (GMP) and Global Wheat Program (GWP), as well as the national program researchers who generated the data used in this and other studies.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendices
Appendix A: Derivation of Full Conditional Distributions for Model BLOR
Liabilities and ω ijt . The fully conditional posterior distribution of liability l ijt is
The last inequality was obtained using a technique called the Pólya-Gamma method (Scott and Pillow 2013), which is useful when working with logistic likelihoods, and has the form
where κ = a − b/2 and P(ω; b, d = 0) denotes the density of the random variable ω ~ PG(b, d = 0), where PG(b, d) denotes a Pólya-Gamma distribution l ijt with parameters b and d and density
\( P\left(\omega; b,d\right)=\left\{{cosh}^b\left(\frac{d}{2}\right)\right\}\frac{2^{b-1}}{\Gamma (b)}\sum_{n=0}^{\infty }{\left(-1\right)}^n\frac{\Gamma \left(n+b\right)\left(2n+b\right)}{\Gamma \left(n+1\right)\sqrt{2\pi {\omega}^3}}\exp \left(-\frac{{\left(2n+b\right)}^2}{8\omega }-\frac{d^2}{2}\omega \right), \)where cosh denotes the hyperbolic cosine.
Then the joint posterior distribution of l ijt and ω ijt is equal to
Therefore, the fully conditional posterior distribution of liability l ijt is a truncated normal distribution and its density is
For simplicity, ELSE is the data and the parameters, except for the one in question. ϕ and Φ are the density and distribution function of a standard normal random variable and the fully conditional posterior distribution l ijt of ω ijt is
From here and from Eq. (4.5) of Polson et al. (2013), we get that
Regression Coefficients ( β )
First note that the fully conditional posterior of l , β , ω is
where \( P\left(\boldsymbol{\omega} \right)={\prod}_{i=1}^I{\prod}_{j=1}^J\prod_{t=1}^{n_{ij}}P\left({\omega}_{ij t};2,0\right) \). Then, the full conditional posterior distribution of β is
where \( {\overset{\sim }{\boldsymbol{\varSigma}}}_0={\left({\boldsymbol{\varSigma}}_0^{-1}{\sigma}_{\beta}^{-2}+{\boldsymbol{X}}^T{\boldsymbol{D}}_{\boldsymbol{\omega}}\boldsymbol{X}\right)}^{-1} \), \( {\overset{\sim }{\boldsymbol{\beta}}}_0={\overset{\sim }{\boldsymbol{\varSigma}}}_0\left({\boldsymbol{\varSigma}}_0^{-1}{\sigma}_{\beta}^{-2}{\boldsymbol{\beta}}_0-{\boldsymbol{X}}^T{\boldsymbol{D}}_{\omega}\left(\sum_{h=1}^2{\boldsymbol{Z}}_h{\boldsymbol{b}}_h\right)+{\boldsymbol{X}}^T{\boldsymbol{D}}_{\omega}\boldsymbol{l}\right) \). It is important to point out that if we use a prior for β ∝ Constant (improper uniform distribution), then in \( {\overset{\sim }{\boldsymbol{\varSigma}}}_0 \) and \( {\overset{\sim }{\boldsymbol{\beta}}}_0 \) we need to make 0 the term \( {\boldsymbol{\varSigma}}_0^{-1}{\sigma}_{\beta}^{-2} \). Finally, the full conditional) posterior of β is
Polygenic effects ( b h )
Now the full conditional posterior of b h is given as
This implies that the full conditional posterior of b h is
with h = 1 , 2, η 1 = Xβ + Z 2 b 2 and η 2 = Xβ + Z 1 b 1.
Variance of polygenic effects \( \left({\sigma}_{b_h}^2\right). \)
Next, the conditional distribution of \( {\sigma}_{b_h}^2 \) is obtained. If \( {\sigma}_{b_h}^2\sim {\chi}^{-2}\left({\nu}_h,{S}_h\right)\left( shape and\ scale\right) \), then
This is the kernel of the scaled inverted χ 2 distribution; therefore, the full conditional posterior is
Threshold effects ( γ )
The density of the full conditional posterior distribution of the cth threshold, γ c , is
If Eq. (4.A.1) is seen as a function of γ c , it is evident that the value of γ c must be larger than all the l ijt |y ijt = c and smaller than all the l ijt |y ijt = c + 1. Hence, as a function of γ c , Eq. (4.A.1) leads to the uniform density
Equation (4.A.2) corresponds to a uniform distribution on the interval [min{min (l ijt |y ijt = c + 1), γ c + 1, γ max }, max{max(l ijt |y ijt = c), γ c − 1, γ min }] (Albert and Chib 1993; Sorensen et al. 1995).
Variance of location effects (\( {\sigma}_{\beta}^2\Big) \)
If we give \( {\sigma}_{\beta}^2\sim {\chi}^{-2}\left({\nu}_{\beta },{S}_{\beta}\right)\left( shape and\ scale\right) \), then
This is the kernel of the scaled inverted χ 2 distribution; therefore, the full conditional posterior is
Appendix B: Derivation of Full Conditional Distributions for Model BNBR
Full conditional for β ∗
where \( {\overset{\sim }{\boldsymbol{\varSigma}}}_0={\left({\boldsymbol{\varSigma}}_0^{-1}{\sigma}_{\beta}^{-2}+{\boldsymbol{X}}^T{\boldsymbol{D}}_{\boldsymbol{\omega}}\boldsymbol{X}\right)}^{-1}, \) \( {\overset{\sim }{\boldsymbol{\beta}}}_0={\overset{\sim }{\boldsymbol{\varSigma}}}_0\left({\boldsymbol{\varSigma}}_0^{-1}{\sigma}_{\beta}^{-2}{\boldsymbol{\beta}}_0-{\boldsymbol{X}}^T{\boldsymbol{D}}_{\omega}\sum_{h=1}^2{\boldsymbol{Z}}_h{\boldsymbol{b}}_h+{\boldsymbol{X}}^T\boldsymbol{\kappa} \right) \).
Full conditional for ω ijt
Full conditional for b 1
Defining η 1 = X β ∗ + Z 2 b 2, the conditional distribution of b 1 is given as
where\( {\boldsymbol{F}}_1={\left({\sigma}_{b_1}^{-2}{\boldsymbol{G}}_1^{-1}+{\boldsymbol{Z}}_1^T{\boldsymbol{D}}_{\omega }{\boldsymbol{Z}}_1\right)}^{-1}\ \mathrm{and}\ {\overset{\sim }{\boldsymbol{b}}}_1={\boldsymbol{F}}_1\left({\boldsymbol{Z}}_1^T\boldsymbol{\kappa} -{\boldsymbol{Z}}_1^T{\boldsymbol{D}}_{\omega }{\boldsymbol{\eta}}^1\right) \).
Full conditional for \( {\sigma}_{b_h}^2 \)
with \( {n}_{b_1}=J \) and \( {n}_{b_2}= IJ \).
Full conditional for \( {\sigma}_{\beta^{\ast}}^2 \)
Full conditional for r
To make the inference of r, we first place a gamma prior on it as r ~ G(a 0, 1/b 0). Then we infer a latent count L for each Y ∼ NB(μ, r) conditional on Y and r. Since L ~ Pois(−r log(1 − π)), by construction we can use the Gamma-Poisson conjugacy to update r. Therefore,
According to Zhou et al. (2012), the conditional posterior distribution of L ijt is a Chinese restaurant table (CRT) count random variable . That is, L ijt ~ CRT(y ijt , r) and we can sample it as \( {L}_{ijt}={\varSigma}_{l=1}^{y_{ijt}}{d}_l, \) where \( {d}_l\sim Bernoulli\left(\frac{r}{l-1+r}\right). \)
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this chapter
Cite this chapter
Montesinos-López, O.A., Montesinos-López, A., Crossa, J. (2017). Bayesian Genomic-Enabled Prediction Models for Ordinal and Count Data. In: Varshney, R., Roorkiwal, M., Sorrells, M. (eds) Genomic Selection for Crop Improvement. Springer, Cham. https://doi.org/10.1007/978-3-319-63170-7_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-63170-7_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-63168-4
Online ISBN: 978-3-319-63170-7
eBook Packages: Biomedical and Life SciencesBiomedical and Life Sciences (R0)