DRaCOoN: Advanced Differential Regulation Analysis for Large-scale Biological Networks

Abstract

Biological complexity necessitates improved methodologies for pinpointing crucial regulatory elements. Existing differential regulation analysis techniques face issues of constrained accuracy and a lack of biological relevance. We introduce DRaCOoN (Differential Regulation and CO-expression Networks), a data-driven method that retrieves differential co-expression and regulatory networks between two distinct conditions. Optimized for large datasets, DRaCOoN embeds algorithmic presentation and benchmarking strategies to counter the limitations of current methods. It calculates several differential metrics and estimates their significance using a permutation test-based approach and a background model. DRaCOoN's internal parallelization capabilities offer faster computation of differential edges, scalability for large datasets, and speed enhancements. We tested DRaCOoN against other methods using a comprehensive simulated benchmark dataset. Our comparative performance analysis demonstrated the superiority of DRaCOoN in various scenarios based on the differential metrics it incorporates. Furthermore, DRaCOoN has successfully spotlighted key regulatory factors in complex biological processes such as bone healing. Overall, our results show that DRaCOoN can be used to monitor mechanisms underlying other complex conditions.