Multi-omics data analysis holds great potential for treatment optimization, molecular diagnostics and disease prognosis. To gain mechanistic insights into diseases, these complex data can be integrated with prior knowledge of molecular interactions and functional relationships using network enrichment methods. The vast majority of existing approaches use pre-determined molecular pathways and thus unable to reveal novel molecular pathways that explain the disease phenotype. In contrast, de novo network enrichment approaches identify disease-associated subnetworks offering mechanistic insights into a disease. However, we observe that these tools are often released in prototype status, drastically limiting the adoption of the tool due to possible software bugs, insufficient documentation and non-intuitive choice of hyperparameters. In the following article, we provide a complete introduction to the field of network enrichment, discuss the algorithmic background of the existing de novo and non-de novo approaches, discuss their limitations and biases and mention possible development perspectives. We present a comprehensive overview of 19 state-of-the-art, well documented and continuously maintained de novo network enrichment tools and offer guidelines for choosing the most appropriate method for a particular analysis.