Biology subjects

Arroyo, J.

Publications and source records attributed to Arroyo, J..

7 recordsLinked to original sources

bioRxiv2026

Phenotypic reversion and target prioritization for cellular inflammation via representation learning with foundation models

The identification of genetic perturbations that can reverse disease-associated cellular phenotypes toward a healthy state is a central challenge in early drug discovery. We present a proof-of-concept framework leveraging single-cell foundation models (scFMs) and a large-scale Perturb-seq dataset to prioritize targets for phenotypic reversion of cellular inflammation. We incorporated both basal and proinflammatory signaling conditions, specifically stimulation with interleukin-1 beta (IL-1{beta}) and tumor necrosis factor alpha (TNF-), to assess whether atherosclerotic disease-relevant stimulation improved identification of genes and pathways critical to disease progression. Our dataset comprised 864,115 endothelial cells subjected to 1,740 unique genetic perturbations. Through having both conditions, we identified targets that exhibited differential effects on gene expression dependent on cellular state. Using scFMs, we embedded single-cell transcriptomes into high-dimensional latent spaces and ranked perturbations by their ability to shift the inflammatory transcriptomic profile toward that of untreated controls. Benchmarking against both annotated gene sets and expert-curated targets, we found robust enrichment for known regulators of inflammation and biologically relevant targets, despite these models having no prior information about these targets. Importantly, including both basal and proinflammatory conditions improved identification of inflammation-associated targets compared to using just the basal condition. This underscored the value of incorporating disease-relevant stimulations in perturbation experiments. Our results highlight the utility of scFMs for data-driven target nomination, emphasize the role of cellular state in regulatory responses, and provide a scalable, model-agnostic approach for ML-guided target discovery. This work offers a valuable community resource for advancing biological understanding of inflammatory-associated disease.

bioinformatics↗

bioRxiv2025

Novel combination of irreversible electroporation and allogenic chimeric antigen receptor (CAR) T-cell therapy synergizes therapeutic outcomes in a preclinical human pancreatic cancer mouse model

Irreversible electroporation (IRE) is a non-thermal ablation modality used clinically for treating unresectable tumors while preserving vital structures through controlled application of pulsed electric fields. Previous data suggest that patient outcomes are enhanced with the induction of an anti-tumor immune response, but current research focuses on using immune checkpoint inhibitors, which function through conventional immune pathways that may be downregulated by cancer or dysregulated by chemo-induced lymphodepletion. Chimeric Antigen Receptor (CAR) T-cells overcome this limitation, as they are engineered with synthetic receptors that redirect lymphocytes to recognize and target cells expressing tumor-specific structures. CARs are engineered to have an increased binding affinity compared to in-situ T-cell binding, amplify internal stimulation cascades, and release pro-inflammatory cytokines that can modulate the endogenous immune system. However, there are still major limitations for adoptive cell therapies in solid tumors, including life-threatening on-target off-tumor cytotoxicity, antigen escape, and failure to infiltrate and persist in solid tumors. Given the substantial evidence that IRE overcomes many of the challenges associated with immune infiltration and persistence in solid tumors, there is a strong premise for using targeted cell therapies following IRE, which would then target residual cancer that could repopulate the lesion. Here, we present the first proof-of-concept combination of IRE with an adoptive cell therapy. We validated that the cell membrane CAR target is not affected in electroporated cells that survive IRE, allowing for subsequent binding and elimination of residual tumor. The research demonstrates the feasibility and synergy of a novel combination of two clinically used techniques.

cancer biology↗

bioRxiv2025

Assessing the Evolutionary Trajectory of Arbuscular Mycorrhizal Conserved Genes in Seagrasses and Aquatic Close Relatives

Arbuscular mycorrhizal fungi (AMF) form beneficial associations with plants, and are thought to have been critical to the adaptation of the ancestor of terrestrial plants during the transition onto land. However, the ability of AMF to associate with aquatic plants is unclear. To address this, we used 65 publicly available genomes and transcriptomes (25 freshwater, 23 terrestrial and 17 marine plants) to interrogate the genomic potential to form AMF associations in aquatic plant lineages in the order Alismatales. We explored the presence or absence of homologs of 45 genes, with a a special focus on six critical genes including three that co-evolved with AMF associations (RAD1, STR1, STR2) and three necessary for intracellular symbiosis (SymRK, CCaMK/DMI3, CYCLOPS/IDP3). Our results indicate a pattern likely consistent with independent gene losses (or extreme divergence) of symbiosis genes across aquatic lineages suggesting a possible inability to form AMF associations. However, some of these conserved genes (i.e., CCaMK/DMI3) are purported to function in other types of fungal symbioses, such as ectomycorrhizal symbiosis, and were observed here in a subset of aquatic lineages, including seagrasses. Overall, our findings highlight the complex evolutionary trajectories of symbiosis-related genes in aquatic plants, suggesting that while AMF associations may have been lost in certain lineages, others have genes that may allow them to form alternative fungal symbioses which may still play an underappreciated role in their ecology.

evolutionary biology↗

bioRxiv2023

Genomic analyses elucidate the causes and consequences of breakdown of distyly in Linum trigynum

Distyly is an iconic floral polymorphism governed by a supergene, which promotes efficient pollen transfer and outcrossing through reciprocal differences in the position of sexual organs in flowers, often coupled with heteromorphic self-incompatibility (SI). Distyly has evolved convergently in multiple flowering plant lineages, but has also broken down repeatedly, often resulting in homostylous, self-compatible populations with elevated rates of self-fertilization. Here, we aimed to study the genetic causes and genomic consequences of the shift to homostyly in Linum trigynum, which is closely related to distylous Linum tenue. Building on a high-quality genome assembly, we show that L. trigynum harbors a genomic region homologous to the dominant haplotype of the distyly supergene conferring long stamens and short styles in L. tenue, suggesting that loss of distyly first occurred in a short-styled individual. In contrast to homostylous Primula and Fagopyrum, L. trigynum harbors no fixed loss-of-function mutations in coding sequences of S-linked distyly candidate genes. Instead, floral gene expression analyses and controlled crosses suggest that mutations downregulating the S-linked LtWDR-44 candidate gene for male SI and/or anther height could underlie homostyly and self-compatibility (SC) in L. trigynum. Population genomic analyses of 224 whole-genome sequences further demonstrate that L. trigynum is highly self-fertilizing, exhibits significantly lower genetic diversity genome-wide, and is experiencing relaxed purifying selection and less frequent positive selection on nonsynonymous mutations relative to L. tenue. Our analyses shed light on the loss of distyly in L. trigynum, and advance our understanding of a common evolutionary transition in flowering plants.

evolutionary biology↗

bioRxiv2022

Genomic analyses of the Linum distyly supergene reveal convergent evolution at the molecular level

Supergenes govern balanced polymorphisms in a wide range of systems. The reciprocal placement of stigmas and anthers in pin and thrum floral morphs of distylous species constitutes an iconic example of a balanced polymorphism governed by a supergene, the distyly S-locus. Recent studies have shown that the Primula and Turnera distyly supergenes are both hemizygous in thrums, but it remains unknown if hemizygosity is pervasive among distyly S-loci. Here we have characterized the genetic architecture and evolution of the distyly supergene in Linum by generating a chromosome-level genome assembly of Linum tenue, followed by the identification of the S-locus using population genomic data. We show that hemizygosity and thrum-specific expression of S-linked genes, including a pistil-expressed candidate gene for style length, are major features of the Linum S-locus. Structural variation is likely instrumental for recombination suppression, and although the non-recombining dominant haplotype has accumulated transposable elements, S-linked genes are not under relaxed purifying selection. Our findings reveal remarkable convergence in the genetic architecture and evolution of independently derived distyly supergenes. The chromosome-level genome assembly and detailed characterization of the distyly S-locus in L. tenue will facilitate elucidation of molecular mechanisms underlying the different forms of flowers described by Darwin.

evolutionary biology↗

bioRxiv2020

Distinct human gut microbial taxonomic signatures uncovered with different sample processing and microbial cell disruption methods for metaproteomic analysis

Metaproteomics is as a promising technique for studying the human gut microbiota, because it can reveal the taxonomic profile and also shed light on the functional role of the microbial community. Nevertheless, methods for extracting proteins from stool samples continue to evolve, in the pursuit of optimal protocols for moistening and dispersing the stool sample and for disrupting microbial cells which are two critical steps for ensuring good protein recovery. Here, we evaluated different stool sample processing and microbial cell disruption methods for metaproteomic analyses of human gut microbiota. An unsupervised principal component analysis showed that different methods produced similar human gut microbial taxonomic profiles. An unsupervised two-way hierarchical clustering analysis identified the microbial taxonomic signatures associated with each method. Proteobacteria and Bacteroidetes identification was favored by moistening the stool samples during processing and by disrupting cells with medium-sized glass beads. Ascomycota identification was enhanced by using large-sized glass beads during sample processing for stool dispersion. Euryarchaeota identification was improved with a combination of small and medium-sized glass beads for cell disruption. Assessments of the relative abundance of Firmicutes, Actinobacteria and Spirochaetes improved when ultrasonication was performed before cell disruption with glass beads. The latter method also increased the overall number of identified proteins. Taxonomic and protein functional analyses of metaproteomic data derived from stool samples from six healthy individuals showed common taxonomic profiles. We also detected certain proteins involved in microbial functions relevant to the host and related mostly to particular taxa, such as B12 biosynthesis and short chain fatty acid production carried out mainly by members in the Prevotella genus and the Firmicutes phylum, respectively. Finally, in this metaproteomic study we identified several human proteins, mostly related to the anti-microbial response, which could contribute to determining the beneficial and detrimental relationships between gut microbiota and human cells in particular human diseases.

microbiology↗

bioRxiv2020

Guiding and interpreting brain network classification with transcriptional data

The investigation of brain networks has yielded many insights that have helped to characterise many neurological and psychiatric disorders. In particular, network classification of functional magnetic resonance imaging (fMRI) data is an important tool for the identification of prognostic and diagnostic biomarkers of brain connectivity disorders such as schizophrenia and depression. However, existing generic network classification methods provide no direct information on the underlying molecular mechanisms of the selected functional connectivity features when applied to fMRI data. To address this, we propose a novel fMRI network classification method that incor-porates brain transcriptional data using a user-specified gene set collection (GSC) to construct feature groups for use in classification of brain connectivity data. The use of GSCs are an opportunity to incorporate knowledge of potential molecular mechanisms which may be associated with a disease. The inclusion of transcriptional data yields improved prediction accuracy on publicly available schizophrenia fMRI data for several of the GSCs we consider. We also introduce a post-hoc interpretation framework to provide transcriptional-data-guided biological interpretations for discriminative functional connectivity features identified by existing fMRI network classification methods.

bioinformatics↗