Abstract
This editorial aims to increase awareness of crucial challenges and limitations in large-scale biomarker discovery models. The focus is on practical data analysis and interpretation aspects that can influence the quality of findings: the prediction capacity and reproducibility of biomarker models. As there is no universal rule to mitigate the effect of potential technological drawbacks, researchers should improve their understanding of possible limitations and pitfalls. Stronger research team proficiency and cooperation are crucial elements in overcoming these obstacles. Moreover, the incorporation of concepts and tools originating from systems science brings new opportunities in terms of predictive power, robustness and interpretability of discoveries.