Biology subjects

Alonso, L.

Publications and source records attributed to Alonso, L..

11 recordsLinked to original sources

bioRxiv2026

Learning from Drops: AI-Guided Integration of Liquid Biopsy Features in Cancer Studies

Cancer is a major global health issue with rising incidence and mortality. Early detection, tumor characterization, and disease surveillance are crucial for timely and effective treatment, ultimately reducing mortality rates. Liquid biopsy (LB) has emerged as a valuable detection tool offering a non-invasive method to determine tumor-derived biomarkers in body fluids with demonstrated translational potential. To increase biomarker sensitivity, high-throughput sequencing platforms deliver massive volumes of data. Artificial Intelligence (AI) is pivotal in enabling huge and complex data integration. This contribution aims to assess the current state of integrative AI-based research in the LB field and provide methodological guidance. First, we conducted a PubMed search and found that the literature is sparse in studies integrating LB features, particularly by applying AI. When adopting the latter approach, defining the study objectives is crucial to guide the subsequent methodological aspects, including study design, patient selection criteria, sample size, nature of the LB features, and metadata to collect. Specifically, we propose strategies and tools for data preprocessing, including normalization and batch correction, as well as handling outliers and missing data. Furthermore, we recommend various Machine/Deep Learning approaches for feature selection techniques to ensure model robustness, and we highlight the importance of undergoing rigorous internal and external validations of the selected models. Assessing clinical utility and interpretability is often overlooked but fundamental for real-world implementation. In conclusion, we provide the LB scientific community with an AI-based methodological guidance to bridge the two fields and enhance the integrative analysis of LB features. Graphical abstractWorkchart for multiomics integrative studies in the liquid biopsy field. Note: CTCs, circulating tumor cells; ctDNA, circulating tumor-DNA; TEPs, tumor-educated platelets; miRNA, microRNA; cfRNAs, cell-free RNAs. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=159 SRC="FIGDIR/small/724535v1_ufig1.gif" ALT="Figure 1"> View larger version (45K): org.highwire.dtl.DTLVardef@1f250b2org.highwire.dtl.DTLVardef@18fe36corg.highwire.dtl.DTLVardef@19c02b9org.highwire.dtl.DTLVardef@176f6e0_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2026

SNMF: Ultrafast, Spatially-Aware Deconvolution for Spatial Transcriptomics

Sequencing-based spatial transcriptomics has revolutionized the study of tissue architecture, but its spots often contain multiple cells, creating a key computational challenge, termed deconvolution, to decipher each spots cell-type composition. Reference-free deconvolution methods avoid the need for a matched single-cell RNA-seq dataset, but typically neglect the spatial correlation between neighboring spots and do not leverage modern hardware for efficient computation. Here, we propose SNMF (Spatial Non-negative Matrix Factorization): a rapid, accurate, and reference-free deconvolution method. SNMF extends the standard NMF framework with a spatial mixing matrix that models neighborhood influences, guiding the factorization toward spatially coherent solutions. Our R package is, to our knowledge, the first spatial transcriptomics deconvolution tool to natively support GPU execution, completing benchmark analyses in under one minute--over two orders of magnitude faster than the slowest competing methods-- with moderate memory requirements. On synthetic and real benchmark datasets, SNMF significantly outperforms state-of-the-art methods in deconvolution accuracy, and on a human melanoma dataset it recovers biologically meaningful cell-type signatures--including a tumor-boundary transition zone-- without any reference input. The proposed mehtod is publicly available at https://github.com/ML4BM-Lab/SNMF.

bioinformatics↗

bioRxiv2025

DEVELOPMENT OF A CONSENSUS MOLECULAR CLASSIFIER FOR PANCREATIC DUCTAL ADENOCARCINOMA

Pancreatic ductal adenocarcinoma (PDAC) presents a significant challenge, with a five-year survival rate of approximately 10%. Tumor heterogeneity contributes to the limited effectiveness of treatments. Several tumor and stroma molecular classifiers have attempted to clarify this heterogeneity with moderate agreement. Recognizing the complexity introduced by this extensive array of taxonomies, this study aims to develop a consensus molecular classifier by including both tumor and stroma features. We integrated gene expression data through Virtual Microdissection and classified the training samples to apply Machine Learning algorithms for each previous classifier. The consensus classifier was then derived using a Markov Clustering Algorithm, and its association with overall survival was assessed. The results indicated that Elastic-Net emerged as the superior model. We identified two classes for tumor components (Consensus Classical and Consensus Non-classical) and stroma components (Consensus Normal-Immune and Consensus Activated-ECM). The consensus Random Forest achieved a balanced accuracy of 96.33% and 98.92%, respectively. While not consistent across retrospective series, the algorithm (PDAConsensus) independently predicted overall survival. We developed a robust consensus classifier for PDAC that integrates tumor and stroma features and made it accessible through the R package PDACMOC (PDACMolecularOmniClassifier, https://github.com/pavillos/PDACMOC) and a Shiny app (https://pdacmoc.cnio.es/).

bioinformatics↗

bioRxiv2024

Secondary motor cortex tracks decision value and supports behavioral flexibility during non-instructed choice

Optimal decision-making relies on interconnected frontal brain regions, which permit animals to adapt their decisions based on their internal state, experience, and environmental context. Among them, the secondary motor cortex (M2) shows earlier decision-related activity required for sensory-guided action selection. However, the role of M2 in adaptive decision-making in the absence of instructive sensory cues remains unclear. Under such conditions, action-selection relies on abstract representations of actions and their values. Using in vivo microscopy and modeling, we showed that M2 neurons in mice exhibited persistent activity encoding decision values (DV) predicting the probability of action-selection during a non-cue-guided lever task. This was confirmed by the reduced reversal performance upon M2 optogenetic inhibition prior to action-selection. Furthermore, updates in DV determined the rate at which learning is reversed. Together, these results provide strong evidence of the use of DV by M2 to adapt choice in the absence of instructive sensory cues. Declaration of interestsThe authors declare no competing financial interests.

neuroscience↗

bioRxiv2024

Serotonin drives aggression and social behaviours of laboratory mice in a semi-natural environment

Aggression is an adaptive social behaviour crucial for the stability and prosperity of social groups. When uncontrolled, aggression leads to pathological violence that disrupts group structure and individual well-being. The comorbidity of uncontrolled aggression across different psychopathologies makes it a potential endophenotype of mental disorders with the same neurobiological substrates. Serotonin plays a critical role in the regulation of impulsive and aggressive behaviours, and mice lacking brain serotonin, due to the ablation of a rate-limiting enzyme of serotonin synthesis (Tryptophan hydroxylase 2, TPH2), are a potential model of pathological aggression. Home cage monitoring allows for the continuous observation and quantification of social and non-social behaviours in group-housed, freely-moving mice. Using an ethological approach, we investigated the impact of central serotonin ablation on everyday expression of social and non-social behaviours and their correlations in undisturbed, group-living Tph2-deficient and wildtype mice. By training a machine learning algorithm on behavioural time series, "allogrooming", "struggling at feeder" and "eating" emerged as key behaviours dissociating one genotype from the other. Although Tph2-deficient mice showed characteristics of pathological aggression and decreased communication compared to wildtype animals they still showed affiliative behaviours to normal levels. Altogether, such distinct and dynamic phenotype of Tph2-deficient mice influenced the groups structure and the dynamic of its hierarchical organization which emerged later. These aspects were analyzed using social network analysis and the Glicko rating methods. This study demonstrates the importance of the ethological approach for understanding the global impact of pathological aggression on different aspects of life both at the individual and the group level. Home cage monitoring allows the observation of the natural behaviours of the mice in a semi-natural habitat and provides an accurate representation of real-world phenomena and pathological mechanisms. The results of this study provide insights into the neurobiological substrate of pathological aggression and their potential role in complex brain disorders.

neuroscience↗

bioRxiv2024

Poor decision making and sociability impairment following central serotonin reduction in inducible TPH2-knockdown rats

Serotonin is an essential neuromodulator for mental health and animals socio-cognitive abilities. However, we previously found that a constitutive depletion of central serotonin did not impair rat cognitive abilities in stand-alone tests. Here we investigated how a mild and acute decrease of brain serotonin would affect rats cognitive abilities. Using a novel rat model of inducible serotonin depletion via the genetic knock-down of tryptophan hydroxylase 2 (TPH2), we achieved 20% decrease of serotonin levels in the hypothalamus after three weeks of non-invasive oral doxycycline administration. Decision making in the Rat Gambling Task and the Probability Discounting Task, cognitive flexibility and social recognition memory were tested in low-serotonin (Tph2-kd) and control rats. Our results showed that the Tph2-kd rats were more prone to choose disadvantageously on the long term (poor decision making) and that only the low-serotonin poor decision makers were more sensitive to probabilistic discounting and had poorer social recognition memory than other low-serotonin and control individuals. Flexibility was unaffected by the acute brain serotonin reduction. Poor social recognition memory was the most central characteristic of the behavioral network of low-serotonin poor decision makers, suggesting a key role of social recognition in expression of their profile. The acute decrease of brain serotonin appeared to specifically amplify cognitive impairments of the subgroup of individuals also identified as poor decision makers in the population. This study highlights the great opportunity Tph2-kd rat model offers to study inter-individual susceptibilities to develop cognitive impairment following mild variations of brain serotonin in otherwise healthy individuals. These transgenic and differential approaches together could be critical for the identification of translational markers and vulnerabilities in the development of mental disorders.

neuroscience↗

bioRxiv2021

Constitutive depletion of brain serotonin differentially affects rats' social and cognitive abilities

BackgroundCentral serotonin is an essential neuromodulator for mental disorders. It appears a promising transdiagnostic marker of distinct psychiatric disorders and a common modulator of some of their key behavioral symptoms. We aimed to identify the behavioral markers of serotonergic function in rats and compare them to human deficits. MethodsWe applied a comprehensive profiling approach in adult male Tph2-/- rats constitutively lacking central serotonin. Under classical and ethological testing conditions, we tested each individuals cognitive, social and non-social abilities and characterized the group organization (i.e. social network, hierarchy). Using unsupervised machine learning, we identified the functions most dependent on central serotonin. ResultsIn classical procedures, Tph2-/- rats presented an unexpected normal cognitive profile. Under the complex and experimenter-free conditions of their home-cage, the same Tph2-/- rats presented drastic changes in their daily life. Brain serotonin depletion induced compulsive aggression and sexual behavior, hyperactive and hypervigilant stereotyped behavior, reduced self-care and body weight, and exacerbated corticosterone levels. Group-housed Tph2-/- rats showed strong social disorganization with disrupted social networks and hierarchical structure, which may arise from communication deficits and cognitive blunting. ConclusionsSerotonin depletion induced a profile reminiscent of the symptomatology of impulse control and anxiety disorders. Serotonin was necessary for behavioral adaptation to dynamic social environments. In classical testing conditions, our animal model challenged the concept of an essential role of serotonin in decision-making, flexibility, and impulsivity, although developmental compensations may have occurred. These contrasting findings highlight the need to generalize the evaluation of animal models multidimensional functions within the complexity of the social living environment.

animal behavior and cognition↗

bioRxiv2021

TIGER: The gene expression regulatory variation landscape of human pancreatic islets

GWAS have identified more than 700 genetic signals associated with type 2 diabetes (T2D). To gain insight into the underlying molecular mechanisms, we created the Translational human pancreatic Islet Genotype tissue-Expression Resource (TIGER), aggregating >500 human islet RNA-seq and genotyping datasets. We imputed genotypes using 4 reference panels and meta-analyzed cohorts to improve coverage of expression quantitative trait loci (eQTL) and developed a method to combine allele-specific expression across samples (cASE). We identified >1 million islet eQTLs (56% novel), of which 53 colocalize with T2D signals (60% novel). Among them, a low-frequency allele that reduces T2D risk by half increases CCND2 expression. We identified 8 novel cASE colocalizations, among which an SLC30A8 T2D associated variant. We make all the data available through the open-access TIGER portal (http://tiger.bsc.es), which represents a comprehensive human islet genomic data resource to elucidate how genetic variation affects islet function and translate this into therapeutic insight and precision medicine for T2D.

bioinformatics↗

bioRxiv2020

A multilayered post-GWAS assessment on genetic susceptibility to pancreatic cancer

Pancreatic cancer (PC) is a complex disease in which both non-genetic and genetic factors interplay. To-date, 40 GWAS hits have been associated with PC risk in individuals of European descent, explaining 4.1% of the phenotypic variance. Here, we complemented a classical new PC GWAS (1D) with spatial autocorrelation analysis (2D) and Hi-C maps (3D) to gain additional insight into the inherited basis of PC. In-silico functional analysis of public genomic information allowed prioritization of potentially relevant candidate variants. We replicated 17/40 previous PC-GWAS hits and identified novel variants with potential biological functions. The spatial autocorrelation approach prioritized low MAF variants not detected by GWAS. These were further expanded via 3D interactions to 54 target regions with high functional relevance. This multi-step strategy, combined with an in-depth in-silico functional analysis, offers a comprehensive approach to advance the study of PC genetic susceptibility and could be applied to other diseases.

genetics↗

bioRxiv2019

Inter-individual and inter-strain differences in cognitive and social abilities of Dark Agouti and Wistar Han rats

BackgroundHealthy animals showing extreme behaviours spontaneously that resemble human psychiatric symptoms are relevant models to study the natural psychobiological processes of maladapted behaviours. Healthy poor decision makers (PDMs) identified using a Rat Gambling Task, co-express a combination of cognitive and reward-based characteristics similar to symptoms observed in human patients with impulse-control disorders. The main goals of this study were to 1) confirm the existence of PDMs and their unique behavioural phenotypes in the Dark Agouti (DA) and Wistar Han (WH), 2) to extend the behavioural profile of the PDMs to probability-based decision-making and social behaviours and 3) to discuss how the key traits of each strain could be relevant for biomedical research.\n\nMethodsWe compared cognitive abilities, natural behaviours and physiological responses in DA and WH rats using several tests. We analysed the results at the strain and the individual level.\n\nResultsPrevious findings in WH rats were reproduced and could be generalized to DA. Each PDM of either strain displayed a similar, naturally occurring, combination of behavioural traits, including possibly higher social rank, but no deficits in probability-based decision-making. A Random forest analysis revealed interesting discriminating traits between WH and DA.\n\nConclusionThe reproducibility and conservation of the socio-cognitive and behavioural phenotypes of GDM (good decision maker) and PDM individuals in the two genetically different strains of WH and DA support a good translational validity of these phenotypes. Both DA and WH rat strains present large phenotypic variations in behaviour pertinent for the study of the underlying mechanisms of poor decision making and associated disorders.

animal behavior and cognition↗

bioRxiv2018

Myc stimulates cell cycle progression through the activation of Cdk1 and phosphorylation of p27

Cell cycle stimulation is a major transforming mechanism of Myc oncoprotein. This is achieved through at least three concomitant mechanisms: upregulation of cyclins and Cdks, downregulation of Cdk inhibitors p15 and p21 and the degradation of p27. The Myc-p27 antagonism has been shown to be relevant in human cancer. To be degraded, p27 must be phosphorylated at Thr-187 to be recognized by Skp2, a component of the ubiquitination complex. We previously described that Myc induces Skp2 expression. Here we show that not only Cdk2 but Cdk1 phosphorylates p27 at the Thr187, which was previously unreported. Moreover, Myc induced p27 degradation in murine fibroblasts through Cdk1 activation, which was achieved by Myc-dependent cyclin A and B induction. In the absence of Cdk2, p27 phosphorylation at Thr-187 was mainly carried out by cyclin A2-Cdk1 and cyclin B1-Cdk1. We also show that Cdk1 inhibition was sufficient for the synthetic lethal interaction with Myc. This result is relevant because Cdk1 is the only Cdk strictly required for cell cycle and the reported synthetic lethal interaction between Cdk1 and Myc.\n\nSummary blurbMyc antagonizes p27 activity in cancer. Myc activates Cdk1 to phosphorylate p27, marking p27 for degradation. This depends on Myc-mediated cyclin A and B induction. Cdk1 inhibition is sufficient for a synthetic lethal interaction with Myc

cancer biology↗