Biology subjects

Caballero, M.

Publications and source records attributed to Caballero, M..

14 recordsLinked to original sources

bioRxiv2026

MET inhibition exploits conserved regulatory dependencies in malignant rhabdoid tumors

Malignant rhabdoid tumors (MRTs) are lethal, yet genomically simple childhood cancers characterised by universal SWI/SNF dysregulation. Despite this shared molecular framework, the regulatory programs sustaining MRTs and how they wire actionable dependencies, remain poorly defined. Here, we integrated single-cell multi-omic profiling and spatial transcriptomics across anatomically and molecularly diverse MRTs, generating a comprehensive multimodal MRT dataset at single-cell resolution. We leveraged this resource to identify candidate tumor-specific dependencies, which were subsequently functionally validated in vitro and in vivo in matched patient-derived models. We found that distinct malignant subpopulations are spatially patterned in complex environmentally shaped topologies, consistent with reciprocal interactions with stromal and immune compartments contributing to intratumoral transcriptional diversity. We identified recurrent transcriptional metaprograms conserved across patients and molecular subtypes, including a previously undescribed shared core rhabdoid program, as well as a metaprogram characterized by extracellular matrix interactions, and preferentially localized in tumor-stroma intermixed neighborhoods, supporting identification of niche conditioning cues. Mechanistically, we identified regulatory wiring driving these metaprograms, with the core shared rhabdoid programme coordinated by a circuit centred on TCF12, ZNF148 and CHD2. Guided by target nominations from our integrated analyses, we performed a curated drug screen using patient-derived MRT models and validated HGF/MET signaling as a selective vulnerability in extrarenal rhabdoid tumors. Pharmacologic MET inhibition was cytotoxic in sensitive models and induced a complete response in vivo using a clinically approved compound. Importantly, MET expression in primary tumors correlated with response, supporting biomarker-guided patient stratification for immediate clinical translation. Collectively, our findings define the molecular architecture of MRTs, reveal conserved regulatory programs beyond anatomical and molecular subtypes, and highlight a targetable signaling axis with immediate translational potential, establishing a framework for precision therapy in this lethal pediatric malignancy.

cancer biology↗

bioRxiv2025

Unraveling the Dual Immunomodulatory and Immunogenic Roles of the Central Conserved Cysteine-Rich Region in Respiratory Syncytial Virus G Protein

Respiratory syncytial virus (RSV) causes severe respiratory disease in infants and high-risk adults, in part by subverting host immunity. The RSV G glycoproteins central conserved cysteine-rich domain (CCD) contains a CX3C motif implicated in immune modulation, but the relationship between CCD redox state, structure and function is unresolved. We recombinantly expressed a CCD peptide (Gpep, residues 149-196) and combined structural and biophysical characterizations with cellular immunology and human serology to define how redox-dependent conformations govern immunogenicity and immunomodulation. Reduced Gpep is compact and rapidly folds via a dominant intermediate into an oxidized, extended monomer; at higher concentrations intermolecular disulfide isomerization produces covalent oligomers. Functionally, monomeric Gpep potently suppresses innate and adaptive activation inhibiting LPS- or UV-inactivated RSV-induced maturation of mouse bone marrow-derived dendritic cells, reducing antigen-specific CD4+ T cell proliferation and IFN-{gamma} production, and attenuating multiple human neutrophil responses (chemotaxis, CD11b upregulation, ROS, MPO release and NET formation), without cytotoxicity. Oligomerized Gpep lacks these suppressive activities. To link molecular mechanism and human exposure, analysis of sera from an ambulatory pediatric cohort (0-72 months) showed a progressive transient increase in the anti-F/anti-G IgG ratio with repeated early RSV exposures, maturating into a functional 2 to 3 ratio. This serologic shift is consistent with previously reported enrichment of F-directed neutralizing immunity. We propose a redox-dependent immune-evasion model in which soluble, monomeric G mediates transient immunosuppression that is removed by disulfide-driven oligomerization, which may occur in membrane-bound G. This may impact therapeutic strategies that commonly favor F-focused responses. ImportanceRespiratory syncytial virus (RSV) remains a leading cause of severe respiratory disease in young children and high-risk adults. This study identifies a chemical switch in a small conserved region of the RSV attachment protein that changes its shape and immune activity: when the region is present as a soluble, single chain, it transiently suppresses key innate and adaptive immune cells, but this activity is lost upon disulfide-mediated oligomerization. We also show that repetitive RSV exposures in early life bias the antibody response, initially boosting antibody responses toward the Fusion protein, rather than this conserved central region. Together, these results may uncover a mechanism by which RSV shapes the host immune response explaining the features of antibody development in children. Understanding this redox-dependent balance between immune evasion and antigenicity will inform safer vaccine and antibody strategies against RSV.

immunology↗

bioRxiv2025

Phenotypic AI-based design of cell-specific small molecule cytotoxics

Drug discovery is being transformed by artificial intelligence, which enables the exploration of vast chemical spaces and the generation of novel compounds with tailored properties. While most generative approaches focus on optimizing physicochemical features or protein-binding affinities, their application to phenotypic drug discovery remains limited. Here we present an integrated framework to design small molecules with selective cytotoxic effects in pancreatic cancer cells that combines high-throughput cytotoxicity screening, bioactivity signature-based machine learning predictors, and reinforcement learning-driven generative models. We screened over 11,000 compounds across six pancreatic cancer and two control cell lines, confirming 392 hits, and used these data to train cell line-specific classifiers that outperformed fingerprint-based models. We embedded these predictors into the REINVENT platform to generate molecules with desired cytotoxicity profiles, including challenging cases of distinguishing between pancreatic cancer lines with highly similar molecular backgrounds. We generated between 37 and 137 high- scoring candidates in seven different cell selectivity exercises, which were structurally distant from the active compounds in the training set. We tested 45 close analogues to the AI-designed molecules, and 20 of them showed the desired cell-specific cytotoxic effects. Indeed, we validated at least one compound in 5 of the 7 exercises tested, showing over a large increase in hit discovery rate compared to the HTS. Additionally, we benchmarked our strategy against direct predictions on a library of over 150,000 compounds, skipping the generative step, finding that the fraction of hits, and more importantly the dose-response validation, is markedly higher in the AI- designed molecules. Overall, our study demonstrates the feasibility of combining predictive and generative AI methods to design molecules with complex phenotypic outcomes in a target- agnostic manner, paving the way for next-generation phenotype-driven drug discovery strategies. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=68 SRC="FIGDIR/small/682546v1_ufig1.gif" ALT="Figure 1"> View larger version (18K): org.highwire.dtl.DTLVardef@119daf2org.highwire.dtl.DTLVardef@7579b9org.highwire.dtl.DTLVardef@8a62f8org.highwire.dtl.DTLVardef@167a6d5_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2022

Welcome to the big leaves: best practices for improving genome annotation in non-model plant genomes

* Premise of the studyRobust standards to evaluate quality and completeness are lacking for eukaryotic structural genome annotation. Genome annotation software is developed with model organisms and does not typically include benchmarking to comprehensively evaluate the quality and accuracy of the final predictions. Plant genomes are particularly challenging with their large genome sizes, abundant transposable elements (TEs), and variable ploidies. This study investigates the impact of genome quality, complexity, sequence read input, and approach on protein-coding gene prediction. * MethodsThe impact of repeat masking, long-read, and short-read inputs, de novo, and genome-guided protein evidence was examined in the context of the popular BRAKER and MAKER workflows for five plant genomes. Annotations were benchmarked for structural traits and sequence similarity. * ResultsBenchmarks that reflect gene structures, reciprocal similarity search alignments, and mono-exonic/multi-exonic gene counts provide a more complete view of annotation accuracy. Transcripts derived from RNA-read alignments alone are not sufficient for genome annotation. Gene prediction workflows that combine evidence-based and ab initio approaches are recommended, and a combination of short and long-reads can improve genome annotation. Adding protein evidence from de novo assemblies, genome-guided transcriptome assemblies, or full-length proteins from OrthoDB generates more putative false positives as implemented in the current workflows. Post-processing with functional and structural filters is highly recommended. * DiscussionWhile annotation of non-model plant genomes remains complex, this study provides recommendations for inputs and methodological approaches. We discuss a set of best practices to generate an optimal plant genome annotation, and present a more robust set of metrics to evaluate the resulting predictions.

bioinformatics↗

bioRxiv2022

Multifactorial heterogeneity of the human mutation landscape related to DNA replication dynamics

Mutations do not occur uniformly across genomes but instead show biased associations with various genomic features, most notably late replication timing. However, it remains contested which mutation types in human cells relate to DNA replication dynamics and to what extents. Previous studies have been limited by the absence of cell-type-specific replication timing profiles and lack of consideration of inter-individual variation. To overcome these limitations, we performed high-resolution comparisons of mutational landscapes between and within lymphoblastoid cell lines from 1662 individuals, 151 chronic lymphocytic leukemia patients, and three colon adenocarcinoma cell lines including two with mismatch repair deficiency. Using cell type-matched replication timing profiles, we demonstrate how mutational pathways can exhibit heterogeneous replication timing associations. We further identified global mutation load as a novel, pervasive determinant of mutational landscape heterogeneity across individuals. Specifically, elevated mutation load corresponded to increased late replication timing bias as well as replicative strand asymmetries of clock-like mutations and off-target somatic hypermutation. The association of somatic hypermutation with DNA replication timing was further influenced by mutational clustering. Considering these multivariate factors, and by incorporating mutation phasing at an unprecedented scale, we identified a unique mutational landscape on the inactive X-chromosome. Overall, we report underappreciated complexity of mutational pathways and their relationship to replication timing and identify specific factors underlying differential mutation landscapes among cell types and individuals.

genetics↗

bioRxiv2022

The evolution of the human DNA replication timing program

DNA is replicated according to a defined spatiotemporal program that is linked to both gene regulation and genome stability. The evolutionary forces that have shaped replication timing programs in eukaryotic species are largely unknown. Here, we studied the molecular causes and consequences of replication timing evolution across 94 humans, 95 chimpanzees, and 23 rhesus macaques. Replication timing differences recapitulated the species phylogenetic tree, suggesting continuous evolution of the DNA replication timing program in primates. Hundreds of genomic regions had significant replication timing variation between humans and chimpanzees, of which 66 showed advances in replication origin firing in humans while 57 were delayed. Genes overlapping these regions displayed correlated changes in expression levels and chromatin structure. Many human-chimpanzee variants also exhibited inter-individual replication timing variation, pointing to ongoing evolution of replication timing at these loci. Association of replication timing variation with genetic variation revealed that DNA sequence evolution can explain replication timing variation between species. Taken together, DNA replication timing shows substantial and ongoing evolution in the human lineage that is driven by sequence alterations and impacts regulatory evolution.

evolutionary biology↗

bioRxiv2021

Comprehensive analysis of DNA replication timing in genetic diseases and gene knockouts identifies MCM10 as a novel regulator of the replication program

Cellular proliferation depends on the accurate and timely replication of the genome. Several genetic diseases are caused by mutations in key DNA replication genes; however, it remains unclear whether these genes influence the normal program of DNA replication timing. Similarly, the factors that regulate DNA replication dynamics are poorly understood. To systematically identify trans-acting modulators of replication timing, we profiled replication in 184 cell lines from three cell types, encompassing 60 different gene knockouts or genetic diseases. Through a rigorous approach that considers the background variability of replication timing, we concluded that most samples displayed normal replication timing. However, mutations in two genes showed consistently abnormal replication timing. The first gene was RIF1, a known modulator of replication timing. The second was MCM10, a highly conserved member of the pre-replication complex. MCM10 mutant cells demonstrated replication timing variability comprising 46% of the genome and at different locations than RIF1 knockouts. Replication timing alterations in MCM10-mutant cells was predominantly comprised of replication initiation defects. Taken together, this study demonstrates the remarkable robustness of the human replication timing program and reveals MCM10 as a novel modulator of DNA replication timing.

genetics↗

bioRxiv2021

Replication timing analysis in polyploid cells reveals Rif1 uses multiple mechanisms to promote underreplication in Drosophila.

Regulation of DNA replication and copy number are necessary to promote genome stability and maintain cell and tissue function. DNA replication is regulated temporally in a process known as replication timing (RT). Rif1 is key regulator of RT and has a critical function in copy number control in polyploid cells. In a previous study (Munden et al., 2018), we demonstrated that Rif1 functions with SUUR to inhibit replication fork progression and promote underreplication (UR) of specific genomic regions. How Rif1-dependent control of RT factors into its ability to promote UR is unknown. By applying a computational approach to measure RT in Drosophila polyploid cells, we show that SUUR and Rif1 have differential roles in controlling UR and RT. Our findings reveal that Rif1 functions both upstream and downstream of SUUR to promote UR. Our work provides new mechanistic insight into the process of UR and its links to RT.

molecular biology↗

bioRxiv2020

Metabolic profile discriminates and predicts Arabidopsis susceptibility to virus under field conditions

As obligatory parasites, plant viruses alter host cellular metabolism. There is a lack of information on the variability of virus-induced metabolic responses among genetically diverse plants in a natural context with daily changing conditions. To decipher the metabolic landscape of plant-virus interactions in a natural setting, one hundred and thirty-two and twenty-six accessions of Arabidopsis thaliana were inoculated with Turnip mosaic virus (TuMV), in two field experiments over 2 years. The accessions were phenotyped for viral accumulation, above-ground biomass, targeted and untargeted metabolic profiles. The accessions revealed quantitative response to the virus, from susceptibility to resistance. Susceptible accessions accumulate primary and secondary metabolites upon infection, at the cost of hindered growth. Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) revealed that the primary metabolites sucrose, glucose and glutamate discriminate susceptible and resistant accessions. Twenty-one metabolic signatures were found to significantly accumulate in resistant accessions whereas they maintained their growth at the same level as mock-inoculated plants without biomass penalty. Metabolic content was demonstrated to discriminate and to be highly predictive of the susceptibility of inoculated Arabidopsis. The PLS coefficient estimated in the training data set reveals, after cross-validation, a correlation of 0.61 between predicted and true viral accumulation. This study is the first to describe the metabolic landscape of plant-virus interactions in a natural setting and its predictive link to susceptibility. It reveals that, in this undomesticated species and in ecologically realistic conditions, growth and resistance are in a permanent conversation and provides new insights on plant-virus interactions.

plant biology↗

bioRxiv2020

The giant sequoia genome and proliferation of disease resistance genes

The giant sequoia (Sequoiadendron giganteum) of California are massive, long-lived trees that grow along the U.S. Sierra Nevada mountains. As they grow primarily in isolated groves within a narrow range, conservation of existing trees has been a national goal for over 150 years. Genomic data are limited in giant sequoia, and the assembly and annotation of the first giant sequoia genome has been an important goal to allow marker development for restoration and management. Using Illumina and Oxford Nanopore sequencing combined with Dovetail chromosome conformation capture libraries, 8.125 Gbp of sequence was assembled into eleven chromosome-scale scaffolds. This giant sequoia assembly represents the first genome sequenced in the Cupressaceae family, and lays a foundation for using genomic tools to aid in giant sequoia conservation and management. Beyond conservation and management applications, the giant sequoia assembly is a resource for answering questions about the life history of this enigmatic and robust species. Here we provide an example by taking an inventory of the large and complex family of NLR type disease resistance genes.

genomics↗

bioRxiv2019

Immune genes, IL1β and Casp9, show sexual dimorphic methylation patterns in the zebrafish gonads

There is a crosstalk between the immune and the reproductive systems in which sexual dimorphism is a common pattern in vertebrates. In the last years, epigenetics has emerged as a way to study the molecular mechanisms involved during gonadal development, which are responsible to integrate environmental information that contributes to assign a specific sexual phenotype (either an ovary or a testis). In the fish gonads, it is known of the existence of the reproduction-immune system interactions although the epigenetic mechanisms involved are far to be elucidated. Here, we used the zebrafish (Danio rerio) as a model to study the DNA methylation patterns of two well-known innate immune genes: IL1{beta} and Casp9. DNA methylation levels were studied by a candidate gene approach at single nucleotide resolution and further, gene expression analysis were carried out. Results showed that there was clear sexual dimorphism in the DNA methylation levels of the two immune studied genes, being significantly higher in the testes when compared to the ovaries. In summary, and although much research is needed, here we present two potential candidates as epimarkers with forthcoming applications in the livestock and fish farming production, for example, in immune fish diseases or sexual control programs.

immunology↗

bioRxiv2019

Comparative Genomics of Six Juglans Species Reveals Patterns of Disease-associated Gene Family Contractions.

Juglans (walnuts), the most speciose genus in the walnut family (Juglandaceae) represents most of the familys commercially valuable fruit and wood-producing trees and includes several species used as rootstock in agriculture for their resistance to various abiotic and biotic stressors. We present the full structural and functional genome annotations of six Juglans species and one outgroup within Juglandaceae (Juglans regia, J. cathayensis, J. hindsii, J. microcarpa, J. nigra, J. sigillata and Pterocarya stenoptera) produced using BRAKER2 semi-unsupervised gene prediction pipeline and additional in-house developed tools. For each annotation, gene predictors were trained using 19 tissue-specific J. regia transcriptomes aligned to the genomes. Additional functional evidence and filters were applied to multiexonic and monoexonic putative genes to yield between 27,000 and 44,000 high-confidence gene models per species. Comparison of gene models to the BUSCO embryophyta dataset suggested that, on average, genome annotation completeness was 89.6%. We utilized these high quality annotations to assess gene family evolution within Juglans and among Juglans and selected Eurosid species, which revealed significant contractions in several gene families in J. hindsii including disease resistance-related Wall-associated Kinase (WAK) and Catharanthus roseus Receptor-like Kinase (CrRLK1L) and others involved in abiotic stress response. Finally, we confirmed an ancient whole genome duplication that took place in a common ancestor of Juglandaceae using site substitution comparative analysis. SIGNIFICANCEHigh-quality full genome annotations for six species of walnut (Juglans) and a wingnut (Pterocarya) outgroup were constructed using semi-unsupervised gene prediction followed by gene model filtering and functional characterization. These annotations represent the most comprehensive set for any hardwood genus to date. Comparative analyses based on the gene models uncovered rapid evolution in multiple gene families related to disease-response and a whole genome duplication in a Juglandaceae common ancestor.

genomics↗

bioRxiv2019

Surprising impacts of crossover interference and sex-specific genetic maps on identical by descent distributions

Simulations of close relatives and identical by descent (IBD) segments are common in genetic studies, yet most past efforts have utilized sex averaged genetic maps and ignored crossover interference, thus omitting features known to affect the breakpoints of IBD segments. We developed Ped-sim, a method for simulating relatives that can utilize either sex-specific or sex averaged genetic maps and also either a model of crossover interference or the traditional Poisson model for inter-crossover distances. To characterize the impact of previously ignored mechanisms, we simulated data for all four combinations of these factors. We found that modeling crossover interference decreases the standard deviation of the IBD proportion by 10.4% on average in full siblings through second cousins. By contrast, sex-specific maps increase this standard deviation by 4.2% on average, and also impact the number of segments relatives share. Most notably, using sex-specific maps, the number of segments half-siblings share is bimodal; and when combined with interference modeling, the probability that sixth cousins have non-zero IBD ranges from 9.0 to 13.1%, depending on the sexes of the individuals through which they are related. We present new analytical results for the distributions of IBD segments under these models and show they match results from simulations. Finally, we compared IBD sharing rates between simulated and real relatives and find that the combination of sex-specific maps and interference modeling most accurately captures IBD rates in real data. Ped-sim is open source and available from https://github.com/williamslab/ped-sim.\n\nAuthor summarySimulations are ubiquitous throughout statistical genetics in order to generate data with known properties, enabling tests of inference methods and analyses of real world processes in settings where experimental data are challenging to collect. Simulating genetic data for relatives in a pedigree requires the synthesis of chromosomes parents transmit to their children. These chromosomes form as a mosaic of a given parents two chromosomes, with the location of switches between the two parental chromosomes known as crossovers. Detailed information about crossover generation based on real data from humans now exists, including the fact that men and women have overall different rates (women produce ~1.6 times more crossovers) and that real crossovers are subject to interference--whereby crossovers are further apart from one another than expected under a model that selects their locations randomly. Our new method, Ped-sim, can simulate pedigree data using these less commonly modeled crossover features, and we used it to evaluate the importance of sex-specific rates and interference in real data. These comparisons show that both factors shape the amount of DNA two relatives share identically, and that their inclusion in models of crossover better fit data from real relatives.

genetics↗

bioRxiv2018

gFACs: Filtering, analysis, and conversion to unify genome annotations across alignment and gene prediction frameworks.

Published genome annotations are filled with erroneous gene models that represent issues associated with frame, start side identification, splice sites, and related structural features. The source of these inconsistencies can often be traced to translated text file formats designed to describe long read alignments and predicted gene structures. The majority of gene prediction frameworks do not provide downstream filtering to remove problematic gene annotations, nor do they represent these annotations in a format consistent with current file standards. In addition, these frameworks lack consideration for functional attributes, such as the presence or absence of protein domains which can be used for gene model validation. To provide oversight to the increasing number of published genome annotations, we present gFACs as a software package to filter, analyze, and convert predicted gene models and alignments. gFACs operates across a wide range of alignment, analysis, and gene prediction software inputs with a flexible framework for defining gene models with reliable structural and functional attributes. gFACs supports common downstream applications, including genome browsers and generates extensive details on the filtering process, including distributions that can be visualized to further assess the proposed gene space.

bioinformatics↗