Symbolic Regression (SR) aims to discover explicit mathematical expressions that describe the relationship between input variables and a target output, offering an interpretable alternative to black‐box machine learning models. Genetic Programming (GP) has been widely adopted for this purpose; however, traditional GP‐based approaches often suffer from high computational cost, limited scalability, and excessive model complexity. To address these limitations, this work presents M5GP 2.0, an extended version of the Multidimensional Multivariate Genetic Programming framework that builds upon the constructive feature induction paradigm introduced in M5GP (Parallel Multidimensional Genetic Programming with Multidimensional Populations for Symbolic Regression). M5GP 2.0 significantly expands the original framework by extending the evolutionary search space through more expressive and higher‐arity operators and introducing advanced GPU‐based optimisations for efficient large‐scale execution. The proposed method evolves symbolic feature transformations that are subsequently combined using linear models, enabling the generation of compact solutions with interpretability potential while maintaining competitive predictive performance. The experimental evaluation is conducted using a standardised and widely accepted benchmark, namely SRBench, enabling rigorous and reproducible comparisons against state‐of‐the‐art methods considering metrics such as predictive R2$$ {R}^2 $$ , root mean squared error (RMSE), model size, and training time. Overall, the results indicate that M5GP 2.0 constitutes a robust and scalable symbolic regression framework that achieves a favourable balance between predictive performance, model compactness, and computational efficiency, while retaining the potential for symbolic traceability and post hoc interpretability.
L. Florido, Leonardo Trujillo, Javier Carmona Troyo et al.· Expert systems· 0 citations
Glial cells, including radial glia, oligodendrocyte precursor cells (OPCs), oligodendrocytes, astrocytes, and microglia, are active and dynamic regulators of central nervous system (CNS) development, homeostasis, and disease. Through extensive interactions with neurons, other glial populations, and the vasculature, they form highly specialized communication networks that are essential for normal brain function. While transcriptomic approaches have revealed extensive glial heterogeneity and enabled the prediction of putative signaling networks, a critical challenge remains in validating and translating these findings at the level of distinct protein complexes existing both within and between the various glial cell types. This is largely due to the fact that traditional proteomic technologies lack spatial resolution and/or fail to capture protein interaction networks. Proximity labeling (PL) has emerged as a powerful strategy to overcome these limitations by enabling cell‐type‐specific mapping of protein networks and subcellular proteomes, with spatial and temporal precision. Emerging studies have applied PL enzymes, such as BioID, TurboID, and HRP, across diverse glial populations, starting to uncover protein networks supporting their interactions with neurons and vascular elements, allowing metabolic support, maintenance of microenvironment homeostasis and cell–cell communication (including synaptic modulation). In this review, we summarize the main PL enzymes, discuss key studies across different glial cell types, and examine the technical challenges and future perspectives of applying PL to investigate glial biology. By complementing transcriptomic data with spatially resolved proteomic insights, PL provides a unique opportunity to deepen our understanding of glial cell biology in health and disease.
João Baltar, R. Abati, L. Florido et al.· Glia· 0 citations
