The increasing volume of unstructured clinical data—such as consultation audio, notes, and reports—has intensified the documentation burden on physicians, limiting time for direct patient interaction. Existing transcription and summarization systems remain largely proprietary and opaque, offering limited semantic interpretability or clinical control. We present AMICO (Anamnesi Medica Intelligente del Colloquio Ospedaliero), a multimodal framework for AI-assisted clinical documentation that integrates audio acquisition, semantic processing, and interactive human validation. The system combines automatic speech recognition, medical named entity recognition, summarization, and drug-drug interaction checking within a unified workflow designed to support, not replace, physician expertise. Evaluation shows that a fine-tuned LLM achieves superior accuracy and faithfulness compared to baseline LLMs, confirming the effectiveness of domain-specific adaptation for Italian medical dialogue summarization. AMICO demonstrates that moderately sized, transparent, and human-supervised models can meaningfully reduce documentation workload while ensuring clinical reliability and safety.

MOTIVATION: In recent years, many algorithms for inferring gene regulatory networks from single-cell transcriptomic data have been published. Several studies have evaluated their accuracy in estimating the presence of an interaction between pairs of genes. However, these benchmarking analyses do not quantify the algorithms' ability to capture structural properties of networks, which are fundamental, e.g., for studying the robustness of a gene network to external perturbations. Here, we devise a three-step benchmarking pipeline called STREAMLINE that quantifies the ability of algorithms to capture topological properties of networks and identify hubs. RESULTS: To this aim, we use data simulated from different types of networks as well as experimental data from three different organisms. We apply our benchmarking pipeline to four inference algorithms and provide guidance on which algorithm should be used depending on the global network property of interest. AVAILABILITY AND IMPLEMENTATION: STREAMLINE is available at https://github.com/ScialdoneLab/STREAMLINE. The data generated in this study are available at https://doi.org/10.5281/zenodo.10710444.

Cells can measure shallow gradients of external signals to initiate and accomplish a migration or a morphogenetic process. Recently, starting from mathematical models like the local-excitation global-inhibition (LEGI) model and with the support of empirical evidence, it has been proposed that cellular communication improves the measurement of an external gradient. However, the mathematical models that have been used have over-simplified geometries (e.g., they are uni-dimensional) or assumptions about cellular communication, which limit the possibility to analyze the gradient sensing ability of more complex cellular systems. Here, we generalize the existing models to study the effects on gradient sensing of cell number, geometry and of long- versus short-range cellular communication in 2D systems representing epithelial tissues. We find that increasing the cell number can be detrimental for gradient sensing when the communication is weak and limited to nearest neighbour cells, while it is beneficial when there is long-range communication. We also find that, with long-range communication, the gradient sensing ability improves for tissues with more disordered geometries; on the other hand, an ordered structure with mostly hexagonal cells is advantageous with nearest neighbour communication. Our results considerably extend the current models of gradient sensing by epithelial tissues, making a step further toward predicting the mechanism of communication and its putative mediator in many biological processes.

MOTIVATION: Dimensionality reduction is a key step in the analysis of single-cell RNA-sequencing data. It produces a low-dimensional embedding for visualization and as a calculation base for downstream analysis. Nonlinear techniques are most suitable to handle the intrinsic complexity of large, heterogeneous single-cell data. However, with no linear relation between gene and embedding coordinate, there is no way to extract the identity of genes driving any cell's position in the low-dimensional embedding, making it difficult to characterize the underlying biological processes. RESULTS: In this article, we introduce the concepts of local and global gene relevance to compute an equivalent of principal component analysis loadings for non-linear low-dimensional embeddings. Global gene relevance identifies drivers of the overall embedding, while local gene relevance identifies those of a defined sub-region. We apply our method to single-cell RNA-seq datasets from different experimental protocols and to different low-dimensional embedding techniques. This shows our method's versatility to identify key genes for a variety of biological processes. AVAILABILITY AND IMPLEMENTATION: To ensure reproducibility and ease of use, our method is released as part of destiny 3.0, a popular R package for building diffusion maps from single-cell transcriptomic data. It is readily available through Bioconductor. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Intercellular communication plays an essential role in multicellular organisms and several algorithms to analyze it from single-cell transcriptional data have been recently published, but the results are often hard to visualize and interpret. RESULTS: We developed Cell cOmmunication exploration with MUltiplex NETworks (COMUNET), a tool that streamlines the interpretation of the results from cell-cell communication analyses. COMUNET uses multiplex networks to represent and cluster all potential communication patterns between cell types. The algorithm also enables the search for specific patterns of communication and can perform comparative analysis between two biological conditions. To exemplify its use, here we apply COMUNET to investigate cell communication patterns in single-cell transcriptomic datasets from mouse embryos and from an acute myeloid leukemia patient at diagnosis and after treatment. AVAILABILITY AND IMPLEMENTATION: Our algorithm is implemented in an R package available from https://github.com/ScialdoneLab/COMUNET, along with all the code to perform the analyses reported here. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

X-Chromosome Inactivation (XCI) is the process whereby one, randomly chosen X becomes transcriptionally silenced in female cells. XCI is governed by the Xic, a locus on the X encompassing an array of genes which interact with each other and with key molecular factors. The mechanism, though, establishing the fate of the X's, and the corresponding alternative modifications of the Xic architecture, is still mysterious. In this study, by use of computer simulations, we explore the scenario where chromatin conformations emerge from its interaction with diffusing molecular factors. Our aim is to understand the physical mechanisms whereby stable, non-random conformations are established on the Xic's, how complex architectural changes are reliably regulated, and how they lead to opposite structures on the two alleles. In particular, comparison against current experimental data indicates that a few key cis-regulatory regions orchestrate the organization of the Xic, and that two major molecular regulators are involved.

At the onset of X-chromosome inactivation, the vital process whereby female mammalian cells equalize X products with respect to males, the X chromosomes are colocalized along their Xic (X-inactivation center) regions. The mechanism inducing recognition and pairing of the X's remains, though, elusive. Starting from recent discoveries on the molecular factors and on the DNA sequences (the so-called "pairing sites") involved, we dissect the mechanical basis of Xic colocalization by using a statistical physics model. We show that soluble DNA-specific binding molecules, such as those experimentally identified, can be indeed sufficient to induce the spontaneous colocalization of the homologous chromosomes but only when their concentration, or chemical affinity, rises above a threshold value as a consequence of a thermodynamic phase transition. We derive the likelihood of pairing and its probability distribution. Chromosome dynamics has two stages: an initial independent Brownian diffusion followed, after a characteristic time scale, by recognition and pairing. Finally, we investigate the effects of DNA deletion/insertions in the region of pairing sites and compare model predictions to available experimental data.