Biology subjects

Flores, M.

Publications and source records attributed to Flores, M..

14 recordsLinked to original sources

bioRxiv2026

PathoBERT: A Hybrid Attention-Based Genomic Language Model for Read-Level Bacterial Pathogenicity Prediction

Motivation Although recent deep learning models have achieved promising results on read level classification tasks, their robustness to realistic sequencing conditions, sensitivity to read length, and ability to support reliable genome level inference remain incompletely characterized. Here, we present PathoBERT, a hybrid deep learning framework that integrates a LoRA adapted DNABERT encoder with convolutional feature extraction, a modified Convolutional Block Attention Module (MCBAM), and Multiscale Convolutional Attention (MSCA) for bacterial pathogenicity prediction. Results The model demonstrated near complete strand invariance and maintained robust performance under simulated sequencing errors, highlighting its suitability for real world next generation sequencing applications. The model operates on individual sequencing reads and supports genome level inference through a read-aggregation strategy based on the Pathogenic Fraction (PathFrac), which combines read level predictions using a majority vote framework. At the read level, PathoBERT outperformed DeePaC at short and moderate fragment lengths (100 and 150 bp), achieving peak performance at 150 bp. At the genome level, PathoBERT achieved perfect separation between pathogenic and non pathogenic genomes using PathFrac based aggregation, resulting in perfect classification performance on the evaluated benchmark compared with competing approaches, namely DeePaC and PathogenFinder 2. Representation level analyses further demonstrated that a progressive refinement of pathogen-associated features throughout the architecture, culminating in highly separable and biologically meaningful latent representations within the final attention pooled embedding space. These findings demonstrate that integrating contextual genomic language models with attention guided multiscale feature extraction provides a robust framework for pathogenicity prediction from short read sequencing data.

bioinformatics↗

bioRxiv2026

Diversity, host-specificity, and environmental drivers of the coral reef eukaryome

Coral reefs are among the most diverse ecosystems on Earth, yet the diversity and structure of their associated microeukaryotic communities remain poorly resolved. We characterized reef-associated eukaryomes across 113 reefs spanning the Pacific Ocean using more than 6,300 samples from corals, seawater, and sediments during the Tara Pacific expedition. We identified [~]121,000 eukaryotic ASVs, revealing one of Earths largest undocumented reservoirs of eukaryotic diversity; over 80% of the recovered diversity was previously undetected in global ocean surveys and fewer than 2% of sequences matched reference databases. Reef habitats supported highly distinct communities, with sediments and seawater harboring 20-40-fold higher richness than corals. Unexpectedly, coral-associated communities across 29 host lineages were consistently dominated by small metazoans, particularly demosponges and maxillopods, identifying these taxa as pervasive and previously unrecognized components of coral eukaryomes alongside Apicomplexa. Across the Pacific, eukaryome composition was strongly structured by environmental gradients, with thermal stress emerging as the primary driver of community turnover. Together, these results identify coral reefs as a globally important reservoir of hidden eukaryotic diversity and reveal the reef eukaryome as a sensitive indicator of ecosystem reorganization under climate change.

systems biology↗

bioRxiv2026

Improving the Accuracy of Forensic Age Estimation Through Bias Reduction

Chronological age estimation can provide supporting information in forensic casework when traditional identification methods are limited. DNA methylation, a stable epigenetic mark, has emerged as a promising tool for predicting chronological age from trace samples. However, many existing age estimation models rely on linear regression approaches, which often yield biased prediction errors across the age distribution (i.e. model residuals show a significant age dependence). In this study, we compared three approaches for age estimation modeling: multivariable linear regression, random forest regression and maximum likelihood estimation. While the first two approaches are well established, for the third one we constructed and validated a DNA methylation-based LOESS regression maximum likelihood model for age estimation utilizing forensic-relevant CpG markers. In all cases, model performance was evaluated through Leave-One-Out Cross-Validation (LOOCV). We utilized three independent publicly accessible methylation datasets collected using droplet digital PCR (ddPCR) to evaluate the most effective method for accuracy and bias in age estimation. Notably, when we compare the results of the maximum likelihood approach to the other approaches, multivariable linear regression and random forest regression, we find less bias in the age associated residuals compared to the other methods. These findings highlight the utility of non-linear modeling techniques in reducing the biases of epigenetic age estimation for forensic applications.

bioinformatics↗

bioRxiv2025

Plasma plasmalogen levels and risk of lymph node positive breast cancer

Lymph node involvement is a key predictor of poor breast cancer prognosis. Systemic lipid alterations can contribute to cancer cell survival in lymph nodes, but their relevance in humans remains unclear. Here, we combine human epidemiologic analyses from the Nurses Health Study 2 and complementary mechanistic studies in mouse models to investigate how systemic lipid profiles relate to lymph node positive breast cancer. We show that in pre-diagnostic human plasma (n=511), lower levels of phosphatidylethanolamine (PE) and phosphatidylcholine (PC)-enriched plasmalogens are associated with increased risk of lymph node positivity, with stronger associations in samples collected closer to diagnosis. Consistent with the human data, lower levels of PE and PC-enriched plasma plasmalogens are found in mice with nodal involvement. Furthermore, in mice, dietary PUFA depletion reduces lipid oxidation in breast cancer cells in lymph nodes and promotes their survival and metastatic spread. These findings suggest that reduced levels of PE and PC-plasmalogens and decreased PUFA availability creates a lipid environment that enables breast cancer lymph node involvement.

cancer biology↗

bioRxiv2025

KSHV Reprograms Host RNA Splicing via FAM50A to Activate STAT3 and Drive Oncogenic Cellular Transformation

RNA alternative splicing is a fundamental cellular process implicated in cancer development. Kaposis sarcoma-associated herpesvirus (KSHV), the etiological agent of multiple human malignancies, including Kaposis sarcoma (KS), remains a significant concern, particularly in AIDS patients. A CRISPR-Cas9 screening of matched primary rat mesenchymal stem cells (MM) and KSHV-transformed MM cells (KMM) identified key splicing factors involved in KSHV-induced cellular transformation. To elucidate the mechanisms by which KSHV-driven splicing reprogramming mediates cellular transformation, we performed transcriptomic sequencing, identifying 131 differential alternative splicing transcripts, with exon skipping as the predominant event. Notably, these transcripts were enriched in vascular permeability, multiple metabolic pathways and ERK1/2 signaling cascades, which play key roles in KSHV-induced oncogenesis. Further analyses of cells infected with KSHV mutants lacking latent genes including vFLIP, vCyclin and viral miRNAs, as well as cells overexpressing LANA, revealed their involvement in alternative splicing regulation. Among the identified splicing factors, FAM50A, a component of the spliceosome complex C, was found to be crucial for KSHV-mediated transformation. FAM50A knockout resulted in distinct splicing profiles in both MM and KMM cells, and significantly inhibited KSHV-driven proliferation, cellular transformation and tumorigenesis. Mechanistically, FAM50A knockout altered SHP2 splicing, promoting an isoform with enhanced enzymatic activity that led to reduced STAT3 Y705 phosphorylation in KMM cells. These findings reveal a novel paradigm in which KSHV hijacks host splicing machinery, specifically FAM50A-mediated SHP2 splicing, to sustain STAT3 activation and drive oncogenic transformation. ImportanceKaposis sarcoma-associated herpesvirus (KSHV) causes cancers such as Kaposis sarcoma, particularly in AIDS patients. This study uncovers how KSHV hijacks a fundamental cellular process called RNA splicing to promote cancer development. We identified key splicing events that alter critical pathways involved in vascular permeability, metabolism, and oncogenic signaling, particularly ERK1/2 and STAT3. A specific protein, FAM50A, was found to be essential for KSHV-driven cancerous transformation. Removing FAM50A disrupted splicing, weakening cancer-promoting signals. These findings provide new insights into how viruses manipulate host cells to drive cancer and highlight RNA splicing as a potential target for future therapies.

cancer biology↗

bioRxiv2023

Decreased accuracy of forensic DNA mixture analysis for groups with lower genetic diversity

Forensic investigation of DNA samples from multiple contributors has become commonplace. These complex analyses use statistical frameworks accounting for multiple levels of uncertainty in allelic contributions from different individuals, particularly for samples containing few molecules of DNA. These methods have been thoroughly tested along some axes of variation, but less attention has been paid to accuracy across human genetic variation. Here, we quantify the accuracy of DNA mixture analysis over 244 human groups. We find higher false inclusion rates for mixtures with more contributors, and for groups with lower genetic diversity. Even for two-contributor mixtures where one contributor is known and the reference group is correctly specified, false inclusion rates are 1e-5 or higher for 56 out of 244 groups. This means that, depending on multiple testing, some false inclusions may be expected. These false positives could be lessened with more selective and conservative use of DNA mixture analysis. HIGHLIGHTSO_LIGroups with lower genetic diversity have higher mixture analysis false positive rates. C_LIO_LIAnalyses with mis-specified references have somewhat higher false positive rates. C_LIO_LIMixture analysis accuracy decreases with more mixture contributors. C_LI GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=198 HEIGHT=200 SRC="FIGDIR/small/554311v2_ufig1.gif" ALT="Figure 1"> View larger version (51K): org.highwire.dtl.DTLVardef@157c28aorg.highwire.dtl.DTLVardef@a4ff05org.highwire.dtl.DTLVardef@62caa0org.highwire.dtl.DTLVardef@1e4ce92_HPS_FORMAT_FIGEXP M_FIG C_FIG

genetics↗

bioRxiv2022

From genome wide SNPs to genomic islands of differentiation: the quest for species diagnostic markers in two scleractinian corals, Pocillopora and Porites

Coral reefs are of paramount importance in marine ecosystems, where they provide support for a large part of the biodiversity. Being quite sensitive to global changes, they are therefore the prime targets for biodiversity conservation policies. However, such conservation goals require accurate species identification, which are notoriously difficult to get in these highly morphologically variable organisms, rich in cryptic species. There is an acute need for easy-to-use and resolutive species diagnostic molecular markers. The present study builds on the huge sequencing effort developed during the TARA Pacific expedition to develop a genotyping strategy to assign coral samples to the correct species within two coral genera (Porites and Pocillopora). For this purpose, we developed a technique that we called "Divergent Fragment" based on the sequencing of a less than 2kb long diagnostic genomic fragment determined from the metagenomic data of a subset of the corals collected. This method has proven to be rapid, resolvable and cost-effective. Sequencing of PCR fragments nested along the species diagnostic fragment allowed us to assign 232 individuals of the genus Pocillopora and 247 individuals of the genus Porites to previously identified independent genetic lineages (i.e. species). This genotyping method will allow to fully analyze the coral samples collected across the Pacific during the Tara Pacific expedition and opens technological perspectives in the field of population genomics-guided conservation.

evolutionary biology↗

bioRxiv2022

Disparate patterns of genetic divergence in three widespread corals across a pan-Pacific environmental gradient highlights species-specific adaptation trajectories.

Tropical coral reefs are among the worst affected ecosystems by climate change with predictions ranging between a 70-90% loss of reefs in the coming decades. Effective conservation strategies that maximize ecosystem resilience, and potential for recovery, must be informed by the accurate characterization of extant genetic diversity and population structure together with an understanding of the adaptive potential of keystone species. Here, we analyzed samples from the Tara Pacific Expedition (2016 to 2018) that completed an 18,000 km longitudinal transect of the Pacific Ocean sampling three widespread corals - Pocillopora meandrina, Porites lobata, and Millepora cf. platyphylla - across 33 sites from 11 islands. Using deep metagenomic sequencing of 269 colonies in conjunction with morphological analyses and climate variability data we can show that the sampled transect encompasses multiple morphologically cryptic species that exhibit disparate biogeographic patterns, and most importantly, distinct evolutionary patterns, despite exposure to identical environmental regimes. Our findings demonstrate on a basin-scale that evolutionary trajectories are species-specific and complex, and can only in part be predicted from the environment. This highlights that conservation strategies must integrate multi-species investigations to consider the distinct genomic footprints shaped by selection as well as the genetic potential for adaptive change.

genetics↗

bioRxiv2021

Viral infection switches the balance between bacterial and eukaryotic recyclers of organic matter during algal blooms

Algal blooms are hotspots of marine primary production and play central roles in microbial ecology and global nutrient cycling. When blooms collapse, organic carbon is transferred to higher trophic levels, microbial respiration or sinking in proportions that depend on the dominant mortality agent. Viral infection can lead to bloom termination, but its impact on the fate of carbon remains an open question. Here, we characterized the consequences of viral infection on the microbiome composition and biogeochemical landscape of marine ecosystems by conducting a large-scale mesocosm experiment. Moniroting of seven induced coccolithophore blooms, which showed different degrees of viral infection, revealed that only high levels of viral infection caused significant shifts in the composition of free-living bacterial and eukaryotic assemblages. Intriguingly, viral infection favored the growth of eukaryotic heterotrophs (thraustochytrids) over bacteria as potential recyclers of organic matter. By combining modeling and quantification of active viral infection at a single-cell resolution, we estimate that viral infection can increase per-cell rates of extracellular carbon release by 2-4.5 fold. This happened via production of acidic polysaccharides and particulate inorganic carbon, two major contributors to carbon sinking into the deep ocean. These results reveal the impact of viral infection on the fate of carbon through microbial recyclers of organic matter in large-scale coccolithophore blooms.

microbiology↗

bioRxiv2021

Developmental and behavioral phenotypes in a new mouse model of DDX3X syndrome

BackgroundMutations in the X-linked gene DDX3X account for ~2% of intellectual disability in females, often co-morbid with behavioral problems, motor deficits, and brain malformations. DDX3X encodes an RNA helicase with emerging functions in corticogenesis and synaptogenesis. MethodsWe generated a Ddx3x haploinsufficient mouse (Ddx3x+/-) with construct validity for DDX3X loss-of-function mutations. We used standardized batteries to assess developmental milestones and adult behaviors, as well as magnetic resonance imaging and immunostaining of cortical projection neurons to capture early postnatal changes in brain development. ResultsDdx3x+/- mice show physical, sensory, and motor delays that evolve into behavioral anomalies in adulthood, including hyperactivity, anxiety-like behaviors, cognitive impairments, and motor deficits. Motor function further declines with age. These behavioral changes are associated with a reduction in brain volume, with some regions (e.g., cortex and amygdala) disproportionally affected. Cortical thinning is accompanied by defective cortical lamination, indicating that Ddx3x regulates the balance of glutamatergic neurons in the developing cortex. ConclusionsThese data shed new light on the developmental mechanisms driving DDX3X syndrome and support face validity of this novel pre-clinical mouse model.

neuroscience↗

bioRxiv2020

Comparative transcriptomic analysis reveals conserved transcriptional programs underpinning organogenesis and reproduction in land plants

The evolution of plant organs, including leaves, stems, roots, and flowers, mediated the explosive radiation of land plants, which shaped the biosphere and allowed the establishment of terrestrial animal life. Furthermore, the fertilization products of angiosperms, seeds serve as the basis for most of our food. The evolution of organs and immobile gametes required the coordinated acquisition of novel gene functions, the co-option of existing genes, and the development of novel regulatory programs. However, our knowledge of these events is limited, as no large-scale analyses of genomic and transcriptomic data have been performed for land plants. To remedy this, we have generated gene expression atlases for various organs and gametes of 10 plant species comprising bryophytes, vascular plants, gymnosperms, and flowering plants. Comparative analysis of the atlases identified hundreds of organ- and gamete-specific gene families and revealed that most of the specific transcriptomes are significantly conserved. Interestingly, the appearance of organ-specific gene families does not coincide with the corresponding organs appearance, suggesting that co-option of existing genes is the main mechanism for evolving new organs. In contrast to female gametes, male gametes showed a high number and conservation of specific genes, suggesting that male reproduction is highly specialized. The expression atlas capturing pollen development revealed numerous transcription factors and kinases essential for pollen biogenesis and function. To provide easy access to the expression atlases and these comparative analyses, we provide an online database, www.evorepro.plant.tools, that allows the exploration of expression profiles, organ-specific genes, phylogenetic trees, co-expression networks, and others.

evolutionary biology↗

bioRxiv2020

A Zinc-dependent metalloproteinase in the intracellular adaptation of Brucella abortus in macrophages

Brucella abortus is a pathogen that survives in macrophages. Several virulence factors participate in this process, including the open reading frame (ORF) BAB1_0270 codifying of a Zinc-dependent metalloproteinase. Here, its contribution in the process of intracellular adaptation was analyzed by infecting RAW264.7 macrophages with the mutant B. abortus {Delta}270 strain. Results showed that this Zinc-dependent metalloproteinase is a cytoplasmic protein that conforms an operon with a transcriptional regulator, which may constitute a type II toxin-antitoxin system. Functionally, this Zinc-dependent metalloproteinase participated neither in the adherence nor the initial intracellular traffic of B. abortus in macrophages. Nevertheless, its deletion significantly increased the co-localization of B. abortus {Delta}270 with phagolysosomal cathepsin D, reducing both its co-localization with calnexin, present in endoplasmic reticulum derived vesicles, and its intracellular replication within macrophages. Besides, B. abortus {Delta}270-infected macrophages produced significantly higher levels of TNF-, IL-6, CD80 and CD86 than B. abortus 2308, even when several genes involved in virulence (vjbR, hutC, bvrR, virB1) were up-regulated in this mutant. Finally, its deletion significantly reduced the capacity of B. abortus {Delta}270 to adapt, grow and express several virulence factors under acidic conditions. Based on these results, we discuss the role of this Zinc-dependent metalloproteinase in the regulation of the virulence of this pathogen, concluding that it contributes significantly to the intracellular adaptation of B. abortus 2308 during the infection of macrophages. Author summaryBrucella abortus is the causative agent of the brucellosis, a highly contagious diseases. A Zinc-dependent metalloproteinase contributes significantly in the intracellular survival. Here, we demonstrate that this metalloproteinase has homology with ImmA/IrrE proteases, which are involved in the bacterial resistance to hostile environment. Furthermore, it conforms a gene pair with a transcriptional regulator, being required by B. abortus to escape from phagolysosomes, to achieve the endoplasmic reticulum and replicate within macrophages. Its deletion from B. abortus stimulated the macrophages, which produced higher levels of pro-inflammatory cytokines and co-stimulatory proteins. This pathogen showed a reduced ability to adapt and grow under acidic conditions, which would negatively affect its escape from phagolysosomes and consequently, stimulating macrophages. Therefore, this work describes how this Zinc-dependent metalloproteinase significantly contributes in the intracellular adaptation of B. abortus 2308 in macrophages.

microbiology↗

bioRxiv2019

Geographical origin determines responses to salinity of Mediterranean caddisflies

Many freshwater ecosystems worldwide, and particularly Mediterranean ones, show increasing levels of salinity. This changes in water conditions could affect abundance and distribution of inhabiting species as well as the provision of ecosystem services. In this study we conduct laboratory experiments using the macroinvertebrate Smicridea annulicornis as a model organism. Our factorial experiments were designed to evaluate the effects of geographical origin on organisms and salinity levels on survival and behavioral responses of caddisflies. The experimental organisms were captured from rivers belonging to three hydrological basins along a 450 Km latitudinal gradient in the Mediterranean region of Chile. Animals were exposed to three conductivity levels, from 180 to 1400 S/cm, close to the historical averages of the source rivers. We measured the behavioral responses to experimental stimulii and the survival time. Our results showed that geography origin shaped the behavioral and survival responses to salinity. In particular, survival and activity decreased more strongly with increasing salinity in organisms coming from more dilute waters. This suggests local adaptation to be determinant for salinity responses in this species of benthic invertebrate. In the current scenario of fast temporal and spatial changes in water levels and salt concentration, the conservation of geographic intra-specific variation of aquatic species is crucial for lowering the risk of salinity-driven biodiversity loss.

ecology↗

bioRxiv2017

Biocuration as an undergraduate training experience: Improving the annotation of the insect vector of Citrus greening disease

The Asian citrus psyllid (Diaphorina citri Kuwayama) is the insect vector of the bacterium Candidatus Liberibacter asiaticus (CLas), the pathogen associated with citrus Huanglongbing (HLB, citrus greening). HLB threatens citrus production worldwide. Suppression or reduction of the insect vector using chemical insecticides has been the primary method to inhibit the spread of citrus greening disease. Accurate structural and functional annotation of the Asian citrus psyllid genome, as well as a clear understanding of the interactions between the insect and CLas, are required for development of new molecular-based HLB control methods. A draft assembly of the D. citri genome has been generated and annotated with automated pipelines. However, knowledge transfer from well-curated reference genomes such as that of Drosophila melanogaster to newly sequenced ones is challenging due to the complexity and diversity of insect genomes. To identify and improve gene models as potential targets for pest control, we manually curated several gene families with a focus on genes that have key functional roles in D. citri biology and CLas interactions. This community effort produced 530 manually curated gene models across developmental, physiological, RNAi regulatory, and immunity-related pathways. As previously shown in the pea aphid, RNAi machinery genes putatively involved in the microRNA pathway have been specifically duplicated. A comprehensive transcriptome enabled us to identify a number of gene families that are either missing or misassembled in the draft genome. In order to develop biocuration as a training experience, we included undergraduate and graduate students from multiple institutions, as well as experienced annotators from the insect genomics research community. The resulting gene set (OGS v1.0) combines both automatically predicted and manually curated gene models. All data are available on https://citrusgreening.org/.

genomics↗