Pedro Carmona-Sáez

Systemic lupus erythematosus (SLE) shows marked clinical and molecular heterogeneity, yet patient stratification often relies on gene expression signatures lacking multicellular context. Here we construct a transcriptional patient map of SLE by analyzing 1,167 total samples (783 SLE, 384 healthy controls) across different resolutions, including single-cell and bulk blood as well as spatially resolved kidney tissue transcriptomes. Using an unsupervised approach we inferred patient-level transcriptomic immune programs from two independent single-cell RNA sequencing cohorts of peripheral blood mononuclear cells (PBMCs), capturing both differences between SLE and health as well as within-SLE heterogeneity. Specifically, we identified four conserved programs comprising two multicellular inflammatory programs driven by interferon and TNF/NFkB activity across immune cells, and two cell type-specific programs reflecting CD8 T cell cytotoxicity and a CD4 T cell naive-to-effector state. Functional analysis of these programs revealed a rewiring of both cell-to-cell interactions and task allocation across cell types during disease activation. In addition, mapping these programs onto an external longitudinal blood transcriptomic cohort predicted flare risk and identified candidate blood protein biomarkers detectable by proteomics. Finally, we showed that these blood programs were enriched in immune-infiltrated glomerular regions from kidney biopsies of individuals with lupus nephritis using spatially resolved transcriptomic data, thereby linking systemic immune programs to local tissue pathology. Graphical

The accelerating development of healthcare data stored in electronic health records (EHRs) has created novel opportunities for biomedical research. Clinical data contains rich, heterogeneous, and longitudinal information of diverse cohorts of population. These data can be analyzed to uncover complex patterns that describe disease progression, comorbidities, and patient trajectories. Significant analytical challenges arise when identifying relationships among clinical variables, often due to high dimensionality and interdependence of temporal data, which requires adaptable methodologies. Classical approaches like Gaussian Graphical Modeling and Vector Autoregression often fall short in addressing these complexities because of the strict required assumptions of independence and stationarity, limiting their applicability to real-world EHRs data. To address this, we present MariNET, a novel approach based on linear mixed models designed to build networks based on clinical variables interactions from longitudinal EHRs. The methodology can effectively handle correlated observations and confounding variables, providing a robust framework for analyzing dynamic interactions between clinical variables over time, offering a scalable and plausible model which is adaptable to researchers’ questions. We show the potential of our method in the analysis of three different datasets. When applied to a COVID-19 mental health cohort, the model successfully captured symptom interactions aligned with previous knowledge. This result demonstrates robustness in handling repeated measures and homogeneous data. In evaluations using Parkinson’s Disease data, we effectively modeled temporal interactions among key clinical variables, aligning with known PD symptom progression and outperforming partial correlation and autoregressive vector models in handling heterogeneous and missing data. In a modified PD dataset by incorporating sex as a random factor, novel method also excelled in correcting biased interactions, unlike partial correlation, which produced misleading results. By using a linear mixed model-based approach, we can effectively capture dynamic interactions in longitudinal EHR data, outperforming traditional methods in analyzing complex clinical datasets. The different use cases here presented support the application of network science in healthcare, enabling more accurate predictions, evidence-based decision-making and improved disease management. The online version contains supplementary material available at 10.1186/s13040-025-00508-y.

Recent studies have emphasized the role of ADAMTS proteases in inflammation and immunity, particularly in the context of tumour progression. Since inflammatory cells can either constrain or promote tumour growth, understanding how ADAMTS proteases influence immune cell behaviour is crucial. Using syngeneic tumour models -B16F1 melanoma and Lewis Lung Carcinoma (LLC)- in Adamts1 knockout (Ats1-KO) mice, we observed model-specific outcomes underscoring the complexity of ADAMTS1 function in cancer. Transcriptomic and functional analyses revealed broad alterations in the matrisome and immune-related pathways across both tumour types. Strikingly, while tumour progression was impaired in B16F1-derived tumours, the LLC model -characterized by a stronger myeloid component- showed no dependency on Adamts1. To investigate this apparent resistance, we experimentally depleted macrophages and uncovered a profound functional defect in this population in Ats1-KO mice. In vitro assays confirmed reduced macrophage phagocytic activity. Mechanistically, we identified the transmembrane heparan sulfate proteoglycan syndecan-4 (SDC4), a known ADAMTS1 substrate, as a key mediator of this activity. Together, these findings reveal a previously unrecognized ADAMTS1-SDC4 axis that links extracellular matrix remodelling to macrophage phagocytosis, ultimately shaping tumour cell clearance and tumour progression.

The proliferation of omics datasets in public repositories has created unprecedented opportunities for biomedical research but has also posed significant challenges for their integration, particularly due to missing genes and platform-specific discrepancies. Traditional gene expression meta-analysis often focuses on individual genes, leading to data loss and limited biological insights when there are missing genes across different studies. To address these limitations, we propose GSEMA (Gene Set Enrichment Meta-Analysis), a novel methodology that leverages single-sample enrichment scoring to aggregate gene expression data into pathway-level matrices. By applying meta-analysis techniques to enrichment scores, GSEMA preserves the magnitude and directionality of effects, enabling the definition of pathway activity across datasets. Using simulated data and case studies on Systemic Lupus Erythematosus (SLE) and Parkinson's Disease (PD), we demonstrate that GSEMA outperforms other methods in controlling false positive rates while providing meaningful biological interpretations. GSEMA methodology is implemented as an R package available on CRAN repository.

Gene set-based single-sample scoring methods are promising to elucidate patient level disease heterogeneity and enable functional interpretation of molecular data for precision medicine approaches. Despite the availability of numerous algorithms, their performance under different scenarios and for downstream applications for precision medicine approaches has not been systematically evaluated. In this study, we conducted a comprehensive survey of an exhaustive list of single-sample scoring methods to assess their stability and reproducibility performances under commo scenarios which include limitations of input data or data integration across studies. We also evaluated their performances for downstream patient stratification and clinical association analyses, as well as predictive modeling of disease states. The in-depth characterization of these scoring methods highlights the importance for a rational design of analysis strategies and provides fundamental insights into method selection under different scenarios or for different applications.

Flow cytometry is an advanced technique for analyzing cellular heterogeneity in biomedical research and clinical diagnostics. Its ability to generate multiparametric data has facilitated advancements in disease classification problems, particularly addressing challenges in distinguishing cell populations and predicting disease outcomes. In these classification tasks, most works consist on either labeling individual cells based on their phenotypic markers or categorizing patient samples as healthy or diseased. However, the complexity of flow cytometry data, characterized by hurdles such as spectral overlap, wide dynamic ranges or batch effects require the usage of preprocessing strategies prior to modeling analysis. This research provides a comprehensive survey of current preprocessing techniques for flow cytometry data used in classification tasks and discusses their specific applications, focusing on four key aspects of data treatment: signal compensation and transformation, batch effect mitigation, imperfect data treatment, and feature selection and class balance. Emphasis is placed on standardizing preprocessing workflows and addressing computational and analytical difficulties posed by the size of modern flow cytometry datasets. The paper also includes a discussion on future opportunities for improving preprocessing pipelines to improve the reproducibility of flow cytometry-based classification models. In short, this work serves as a reference for experts in the field, consolidating best practices in preprocessing and providing guidelines for the development of methodologies that optimize the classification of flow cytometry data.

Introduction: Alopecia areata (AA) is an autoimmune disorder characterized by non-scarring hair loss due to immune dysregulation. Despite advances, its precise molecular mechanisms remain unclear. This study investigates plasma miRNA expression profiles in patients with AA to identify biological pathways influenced by miRNAs and potential therapeutic targets.Methods: Fifty patients with AA were categorized as severe or mild based on SALT scores. Plasma miRNA levels were compared with those of healthy controls and individuals with other immune-mediated skin diseases. In the discovery phase, 754 miRNAs were analyzed in 20 participants (five severe AA, five mild AA, and ten controls). Key miRNAs identified were then validated in a second cohort of 90 participants, including patients with AA, non-segmental vitiligo, atopic dermatitis (AD), psoriasis (PsO), and healthy controls, using RT-PCR. Machine learning was used to classify patients based on their miRNA profiles, and pathway enrichment analysis and drug targeting were conducted to explore therapeutic opportunities.Results: Nineteen miRNAs were significantly downregulated in AA, with nine technically and clinically validated for both mild and severe forms. The top four miRNAs with the highest classification potential were miR-130b-3p, miR-296-5p, miR-424-5p, and miR-195-5p. Distinct upregulation patterns were identified in vitiligo, AD, and PsO. Machine learning models showed vital classification accuracy for AA (AUC = 0.94) and PsO (AUC = 0.88), with moderate performance for vitiligo and AD. Pathway enrichment analysis highlighted immune-related pathways, including the Interferon-gamma and JAK/STAT signaling pathways. Drug repositioning identified kinase inhibitors showing the most significant promise for reversing miRNA dysregulation.Conclusion: This study identifies distinct plasma miRNA profiles in AA, with potential applications for both diagnosis and therapy. Machine learning validated its solid predictive accuracy, and pathway analysis highlighted key immune pathways in AA. These findings should be interpreted as exploratory and hypothesis-generating, pending further functional validation of candidate miRNAs.

Recent advances in single-cell RNA-Seq (scRNA-Seq) technologies have revolutionized our ability to gather molecular insights into different phenotypes, such as diseases, at the level of individual cells. The analysis of the resulting data poses significant challenges due to their sparsity and large volume, and proper statistical methods are required to analyze and extract information from scRNA-Seq datasets. Sample classification based on gene expression data has proven effective and valuable for precision medicine applications. However, standard classification schemas are often not suitable for scRNA-Seq due to its unique characteristics, and new algorithms are required to effectively analyze and classify samples at the single-cell level. In this article, we introduce singleDeep, an end-to-end pipeline that streamlines the analysis of scRNA-Seq data training deep neural networks, enabling robust prediction and characterization of sample phenotypes. To validate the effectiveness of singleDeep, we applied it to make predictions on scRNA-Seq datasets from different conditions, including systemic lupus erythematosus and Alzheimer's disease. Our results demonstrate strong diagnostic performance, validated both internally and externally. Moreover, compared with traditional machine learning methods applied to pseudobulk data, singleDeep consistently outperformed these approaches. In addition to prediction accuracy, singleDeep provides valuable insights into cell types and gene importance estimation for phenotypic characterization. This functionality provided additional and valuable information in our use cases. For instance, we corroborated that some interferon signature genes are consistently relevant for autoimmunity across all immune cell types in lupus. On the other hand, we discovered that genes linked to dementia have relevant roles in specific brain cell populations, such as APOE in astrocytes.

Recent advances in single-cell RNA-Sequencing (scRNA-Seq) technologies have revolutionized our ability to gather molecular insights into different phenotypes at the level of individual cells. The analysis of the resulting data poses significant challenges, and proper statistical methods are required to analyze and extract information from scRNA-Seq datasets. Sample classification based on gene expression data has proven effective and valuable for precision medicine applications. However, standard classification schemas are often not suitable for scRNA-Seq due to their unique characteristics, and new algorithms are required to effectively analyze and classify samples at the single-cell level. Furthermore, existing methods for this purpose have limitations in their usability. Those reasons motivated us to develop singleDeep, an end-to-end pipeline that streamlines the analysis of scRNA-Seq data training deep neural networks, enabling robust prediction and characterization of sample phenotypes. We used singleDeep to make predictions on scRNA-Seq datasets from different conditions, including systemic lupus erythematosus, Alzheimer’s disease and coronavirus disease 2019. Our results demonstrate strong diagnostic performance, validated both internally and externally. Moreover, singleDeep outperformed traditional machine learning methods and alternative single-cell approaches. In addition to prediction accuracy, singleDeep provides valuable insights into cell types and gene importance estimation for phenotypic characterization. This functionality provided additional and valuable information in our use cases. For instance, we corroborated that some interferon signature genes are consistently relevant for autoimmunity across all immune cell types in lupus. On the other hand, we discovered that genes linked to dementia have relevant roles in specific brain cell populations, such as APOE in astrocytes.

Accurately predicting patient outcomes is essential for optimizing treatment and improving outcomes in pediatric acute myeloid leukemia (AML). In recent years, microRNAs have emerged as a promising prognostic marker, with a growing body of evidence supporting their potential predictive value. We systematically reviewed all previous studies that have analyzed the expression of microRNAs as predictors of survival in pediatric AML and found 16 microRNAs and 4 microRNA signatures previously proposed as predictors of survival. We then used a public access cohort of 1414 pediatric AML patients from the TARGET project to develop a new predictive model using penalized lasso Cox regression based on microRNA expression. Here we propose a new score based on a 37-microRNA signature that is associated with AML and is able to predict survival more accurately than previous microRNA-based methods.