Computational workflows typically consist of many tools that are usually distributed as compiled binaries or source code. Each of these software tools typically depends on other installed software, and performance could potentially vary due to versions, updates, and operating systems. We show here that the analysis of mRNA-seq data can depend on the computing environment, and we demonstrate that software containers represent practical solutions that ensure the reproducibility of RNAseq data analyses.