High-dimensional data, with many more covariates than observations, such as genomic data for example, are now commonly analyzed. However, there are few tools for high-dimensional variable selection when the data are observations collected repeatedly on several individuals, and even fewer when the model is nonlinear. Thus, we develop a high-dimensional covariate selection procedure for nonlinear mixed-effects models that are natural models for analyzing this type of data. More precisely, we propose a spike-and-slab variable selection in which we fit using the stochastic approximation version of EM algorithm. Similarly to lasso regression, the set of relevant covariates is selected by exploring a grid of values for the penalization parameter. The proposed approach is much faster than a classical MCMC algorithm and shows very good selection performances on simulated data. The efficiency of the proposed method is illustrated on a problem of genetic markers identification, relevant for genomic assisted selection in plant breeding. The current aim is to achieve consistency in model selection for this problem, which is a work in progress.