Network estimation and variable selection have been extensively studied in\nthe statistical literature, but only recently have those two challenges been\naddressed simultaneously. In this paper, we seek to develop a novel method to\nsimultaneously estimate network interactions and associations to relevant\ncovariates for count data, and specifically for compositional data, which have\na fixed sum constraint. We use a hierarchical Bayesian model with latent layers\nand employ spike-and-slab priors for both edge and covariate selection. For\nposterior inference, we develop a novel variational inference scheme with an\nexpectation maximization step, to enable efficient estimation. Through\nsimulation studies, we demonstrate that the proposed model outperforms existing\nmethods in its accuracy of network recovery. We show the practical utility of\nour model via an application to microbiome data. The human microbiome has been\nshown to contribute to many of the functions of the human body, and also to be\nlinked with a number of diseases. In our application, we seek to better\nunderstand the interaction between microbes and relevant covariates, as well as\nthe interaction of microbes with each other. We provide a Python implementation\nof our algorithm, called SINC (Simultaneous Inference for Networks and\nCovariates), available online.\n