The doctoral thesis presents a unified computational framework for extracting robust and biologically meaningful knowledge from high-dimensional biomedical data, structured around two complementary research lines.

The first research line focuses on consensus inference of Gene Regulatory Networks (GRNs). Throughout the thesis, several methods were proposed, including Single-GENECI [1], Memetic Infer [2], MO-GENECI [3], PBEvoGen [4], and BIO-INSIGHT [5], demonstrating progressive improvements in inference accuracy on simulated benchmarks. These approaches were subsequently applied to real patient data from melanoma, fibromyalgia, and myalgic encephalomyelitis. All these developments were consolidated into the open-source software package GENECI, available on GitHub and PyPI, with over 17,000 downloads.

The second research line addresses the biclustering of biomedical data through evolutionary optimization. This work resulted in MOEBA-BIO [6], a generic evolutionary biclustering framework for biomedical applications. Based on the knowledge acquired in the GRN inference line, the framework was later specialized into MOEBA-BIO-CoExp [6] for gene co-expression analysis by introducing an objective function that enforces coherence between biclusters and regulatory communities detected in GRNs inferred from the input data.