The following article considers a mixture of regressions with variable selection problem. In many real-data scenarios, one is faced with data which possess outliers, skewness and, simultaneously, one would like to be able to construct clusters with specific predictors that are fairly sparse. A Bayesian mixture of lasso regressions with t-errors to reflect these specific demands is developed. The resulting model is necessarily complex and to fit the model to real data, a state-of-the-art Particle Markov chain Monte Carlo (PMCMC) algorithm based upon sequential Monte Carlo (SMC) methods is developed. The model and algorithm are investigated on both simulated and real data. © 2014 Elsevier B.V. All rights reserved.