The Impact of Pathway Database Choice on Statistical Enrichment Analysis and Predictive Modeling

BackgroundPathway-centric approaches are widely used to interpret and contextualize -omics data. However, databases contain different representations of the same biological pathway, which may lead to different results of statistical enrichment analysis and predictive models in the context of precision medicine.\n\nResultsWe have performed an in-depth benchmarking of the impact of pathway database choice on statistical enrichment analysis and predictive modeling. We analyzed five cancer datasets using three major pathway databases and developed an approach to merge several databases into a single integrative database: MPath. Our results show that equivalent pathways from different databases yield disparate results in statistical enrichment analysis. Moreover, we observed a significant dataset-dependent impact on performance of machine learning models on different prediction tasks. Further, MPath significantly improved prediction performance and reduced the variance of prediction performances in some cases. At the same time, MPath yielded more consistent and biologically plausible results in the statistical enrichment analyses. Finally, we implemented a software package designed to make our comparative analysis with these and additional databases fully reproducible and to facilitate the update of our integrative pathway resource in the future.\n\nConclusionThis benchmarking study demonstrates that pathway database choice can influence the results of statistical enrichment analysis and prediction modeling. Therefore, we recommend the use of multiple pathway databases or the use of integrative databases.