Search bioRxiv⌕ Search

Biology subjects

Ahmad, S.

Publications and source records attributed to Ahmad, S..

At least 19 recordsLinked to original sources

Mechanism-guided Discovery of Fluoroacetate Dehalogenase (FAcD) Variants for Enzymatic PFAS Defluorination

Per- and polyfluoroalkyl substances (PFAS) are persistent anthropogenic "forever chemicals" whose strong C-F bonds (~460-533 kJ mol-1) resist defluorination, and the recast EU Drinking Water Directive sets parametric values of 0.10 ug L-1 for the sum of 20 specified PFAS and 0.50 ug L-1 for total PFAS. Here, we present DeFluorX, a mechanism-first pipeline that uses Boltz-2 to co-fold 2,150 bacterial fluoroacetate dehalogenase (FAcD) variants with 26 PFAS and fluoroacetate, generating uniformly high model confidence (rho = +0.08) but poorly tracked catalytic competence. Thus, the complexes were ranked and tiered by SN2-defluorination geometry based on the known mechanism of control of FAcDs RPA1163 with fluoroacetate: Odelta...C-alpha distance and Odelta-C-alpha-F angle; relay; carboxylate clamp; F-stabiliser; F-cradle; and penalties for scissile bond strength, backside occlusion, beta-fluorination and bond multiplicity. The tiering identified eight complexes (med angle 173.2 degrees, distance 2.91 A) with a high probability of degradation, including Alphaproteobacterial enzymes, with mono-, di- and trifluoroacetate attacking at 172-178 degrees versus 163 degrees for the control. We performed WaterMap, restrained 1 us Desmond MD, MM-GBSA and scans of the three WaterMap-aware best MD frames per complex using B3LYP/6-31G** QM/MM. All 12 frames exhibited product-end C-F cleavage with barriers ranging from 1.1 to 11.8 kcal mol-1, and the departing fluoride carried -0.43 to -0.63 e. The Laboratory validation is ongoing.

bioinformatics↗

A staged evidence chain identifies a recent horizontal transfer of a HET-domain gene into a Fusarium tricinctum strain from a lineage related to the Fusarium oxysporum species complex

Genes encoding the heterokaryon incompatibility (HET) domain form fast-evolving, highly paralogous families, which makes horizontal gene transfer (HGT) among them hard to separate from duplication, loss and annotation artefacts. We describe a staged evidence chain for testing a single HGT candidate and apply it to the strongest signal in the HET-domain view of a 109-genome similarity graph of 8,815 HET-domain (PF06985) proteins. Each stage is a separate test with a decision rule fixed before it was run: (i) a gene tree with ultrafast-bootstrap support; (ii) a species-tree check that donor and recipient lie in different clades; (iii) a genome-wide identity background and a contamination screen of the recipient assembly; (iv) a sister-species control that removes genes conserved by vertical descent; and (v) a tblastn search of whole genome assemblies, so that absence is tested in the DNA rather than in the gene annotation. We identify a recent horizontal transfer of a HET-domain gene into one F. tricinctum strain (MPI-SDFR-AT-0068) from a lineage related to the F. oxysporum species complex (FOSC). The F. tricinctum copy is 98.4% identical to its best FOSC hit (local protein identity), against a genome-wide median of 83%, and its synonymous divergence from the FOSC copies (median dS 0.11) is lower than that of 99.6% of F. tricinctum genes from their best F. oxysporum hits and of 91% of F. tricinctum genes from their best hits in the sister species, which argues strongly against vertical descent, hidden paralogy and shared ancestral alleles. Protein gene trees group it with the FOSC copies (UFBoot 96-98), and a nucleotide tree of the gene region places it as sister to all sampled FOSC copies (UFBoot 98). F. tricinctum and the F. oxysporum strains belong to different species complexes (species-tree patristic distance 0.075, about 22 times the distance between F. oxysporum strains). The gene is absent from the assembly of F. tricinctum's sister species, F. avenaceum (best locus 49% identity), and from 14 of 15 further F. avenaceum genomes, from 23 of the 26 assemblies of Fusarium species outside the FOSC and the F. tricinctum complex, and no other sequenced genome of the recipient strain's own lineage carries it; close copies (at least 90% identity) occur only in 7 of 9 FOSC strains and in one metagenome-assembled Fusarium genome from neither complex. In F. tricinctum the gene sits in an approximately 6-kb strain-specific island on a putative accessory scaffold with F. oxysporum-affine gene content, whereas in the FOSC it occupies one conserved, subtelomeric locus of a core chromosome. Assembly contamination was excluded, and single long reads span the gene and its flanks. The donor lineage remains unresolved, and formal reconciliation is still needed. The evidence chain is cheap and uses only public assemblies; each stage targets a specific alternative explanation, and synonymous divergence against the genome background proved the most decisive test.

genomics↗

TFA-defluorinating microbial consortium achieved with analogue enrichment

Trifluoroacetate (TFA) is an ultra-short chain perfluoroalkyl substance (PFAS) and a pollutant of global significance; it is persistent and ubiquitous in the environment across compartments, including food, drinking water, animals, and humans. Concentrations in the environment are rising, concerns over the toxicity of the compound are increasing, and there are currently no efficient ways to remediate TFA pollution, as conventional methods are ineffective at scale under environmental conditions. Microorganisms possessing dehalogenating enzymes targeting the structurally analogous compound monofluoroacetate (MFA) present a potential solution to degrade TFA microbiologically. Due to previously demonstrated dehalogenation of difluoroacetate (DFA) by dehalogenases, the promiscuity of such enzymes was hypothesized to be extendable to TFA as well. In this study cultures obtained from an environmental sample of PFAS contaminated wastewater sludge were incubated in M9 minimal media supplied with 33 mM acetate and either 33 mM TFA, DFA, or MFA in an iterative manner, followed by testing in a 30-day experiment. Free fluoride (F-) was measured at 5 timepoints using a fluoride ion specific electrode (ISE). One culture designated MFA1C significantly degraded MFA, DFA, and TFA (two-way ANOVA mixed model, p < 0.05). MFA1C released 11.2 {+/-} 0.7 M F- which constitutes approximately 0.01% of the potential fluoride release from the total 33 mM TFA applied and with the greatest defluorination activity in the first 9 days of incubation. Fluoride release from TFA and formation of DFA and MFA was confirmed with 19F-NMR. Genetic sequencing and protein-ligand modelling identified two candidate dehalogenase enzymes from Achromobacter denitrificans and Stenotrophomonas sp., respectively, with potential capacity to degrade TFA. The findings of this study are a crucial first step towards future potential bioremediation of TFA, and are, to the best of our knowledge, the first to identify a TFA-degrading consortium of microorganisms.

microbiology↗

Cross-cohort Mamography]{Representation-dependent domain adaptation for cross-cohort generalization of mammography diagnostic models

Background: Artificial intelligence (AI) models for mammography can achieve high diagnostic performance when training and test data originate from similar populations and imaging environments, but performance often deteriorates under cross- cohort domain shift. Differences in acquisition systems, image characteristics, preprocessing, population composition, and disease prevalence can introduce distributional variation that is poorly captured by conventional within- dataset validation. Recent studies have further shown that Foundation Model representations can retain substantial dataset-specific information. However, it remains unclear how image representation influences diagnostic transfer to completely unseen cohorts and whether domain-adversarial learning provides comparable benefits across different representation types. Methods: We systematically evaluated cross-cohort mammography classification using seven independent datasets spanning diverse geographic, institutional, scanner, and imaging environments. Three image representations were compared: handcrafted radiomic features, EfficientNet-B0-derived features, and embeddings from the medical vision-language model MedSigLIP-448. Each was evaluated under three preprocessing conditions: unprocessed images, Otsu-based breast-region extraction, and Otsu extraction combined with Contrast Limited Adaptive Histogram Equalization (CLAHE). Patient-level classification was performed using a mean-pooled multilayer perceptron (MLP) baseline and an attention-based Domain-Adversarial Neural Network (DANN). Generalization was evaluated using a multi-source Leave-One-Domain-Out (LODO) framework in which six datasets were used for model development and the seventh was completely withheld as an unseen test cohort, iteratively across all seven datasets. Performance was primarily assessed by area under the receiver operating characteristic curve (AUC-ROC). Results: Cross-cohort performance depended strongly on image representation and model architecture. MedSigLIP consistently provided the highest performance on unseen cohorts. With MedSigLIP features, mean LODO AUC-ROC increased from 0.598 with MLP to 0.685 with DANN for unprocessed images, from 0.607 to 0.678 after Otsu preprocessing, and from 0.603 to 0.659 after CLAHE+Otsu preprocessing. When DANN-MLP differences were evaluated across held-out datasets and preprocessing conditions, a significant gain was observed for MedSigLIP after correction for multiple comparisons, whereas corresponding gains were not significant for Radiomics or EfficientNet-B0. A mixed-effects analysis similarly showed a positive MedSigLIP X DANN interaction, although this narrowly missed conventional statistical significance. Unprocessed images produced the highest numerical mean performance for MedSigLIP-DANN, but preprocessing itself had no significant overall effect. Notably, linear probing showed strong dataset discriminability in MedSigLIP and EfficientNet representations but little linearly recoverable dataset information in Radiomics, indicating that low dataset separability alone did not predict better diagnostic generalization. Conclusion: Cross-cohort mammography generalization depends strongly on the choice of image representation and its interaction with the classification strategy. MedSigLIP provided the most transferable representation in our multi-source unseen-domain evaluation, and the performance gain associated with DANN was concentrated in MedSigLIP rather than consistently observed across representation types. These findings suggest that the effectiveness of domain-adversarial learning is representation-dependent and that neither conventional preprocessing nor low dataset discriminability alone ensures robust transfer to unseen clinical cohorts.

bioinformatics↗

Microbes on the Mind: Multi-modal Neuroimaging Reveals Gut-Brain Axes in Infants

The gut microbiome develops from birth and is associated with cognition and mental health, but its relationship with early brain maturation remains unclear. We studied 223 typically developing infants aged 0-3 years using multimodal MRI and paired fecal shotgun metagenomics. Microbiome-wide association analysis linked microbial features to cortical morphology, white-matter microstructure, and functional networks. Regularized canonical correlation analysis identified structural-metabolic and functional-taxonomic covariance patterns. Sparse partial least-squares analysis recapitulated both full-sample CCA patterns and their defining structural and functional phenotypes, providing cross-method corroboration. Longitudinal canonical-score slopes did not significantly covary, whereas earlier neuroimaging CCA1 scores predicted microbiome CCA1 scores 1-6 months later in bidirectional autoregressive cross-lag models. These observational results identify coordinated developmental patterns and temporal ordering without establishing causality.

developmental biology↗

Chemically induced USP21-KLHL12 association recruits USP21 to COPII assembly

Ubiquitin-specific protease 21 (USP21) is a deubiquitinase implicated in various biological processes and disease states, including cancer. To better understand USP21 protein interaction networks, we used biotin identification proximity labeling, and uncovered USP21 interactions with the cytoskeleton, endosomal, and transcriptional regulators. Notably, a subset of these interactions was modulated by the USP21 selective inhibitor BAY-805. In cells, BAY-805 promoted proximity between USP21 and Kelch like protein KLHL12, a component of the Cullin3 ligase complex. Structural modeling using AlphaFold3, together with hydrogen deuterium exchange mass spectrometry, identified an orthosteric BAY-805 binding site within USP21. Binding of BAY-805 functionally competes with ubiquitin at USP21, thereby enabling USP21 interaction with KLHL12. This BAY-805 induced proximity elicited relocalization of USP21 to KLHL12 containing coat protein complex II (COPII) vesicles. Induced USP21-KLHL12 association led to enlargement of COPII vesicles and impaired transport of large cargoes such as collagen. Overall, our findings reveal a novel chemically induced proximity mechanism in which the monovalent ligand BAY-805 outcompetes ubiquitin from USP21, thereby recruiting USP21 to KLHL12 associated protein complexes.

cell biology↗

Evidence for early kidney vasculature development by vasculogenesis and angiogenesis based on advanced in vivo chimeric model and depletion of CD31+/CD146+ intrinsic endothelial progenitors

Kidney diseases are a significant health concern worldwide, with a growing demand for fundamental treatments, such as generation of transplantable organs from programmed cells. The generation of kidney organoids has led to significant advances in disease modeling. However, failure to establish a robust endothelial vasculature network in kidney organoids remains a barrier to oxygen support and gas exchange in renal tissue engineering. We targeted the development of early kidney vasculature system by advancing the classic mouse embryonic kidney organ culture and organoid model system by introducing a novel mini reservoir setup. The method improved integration with the highly vascularized chicken Chorion Allantois Membrane (CAM), enabling host blood flow to developing kidney organoids. We examined the origin and role of the CD31+ endothelial cells (ECs) and their CD146+ progenitors by applying a dissociated embryonic kidney approach. This enabled selected depletion of specific cell populations, followed by regeneration of the kidney organ primordia by reaggregation and placement on the CAM to assess organogenesis. The cell lineage and fate tracing with constitutively labelled cells revealed that the intrinsic endothelial CD146+ progenitor and those of CD31+ ECs in the kidney initiating organogenesis established the kidney vasculature. If the intrinsic EC progenitors were depleted, the mouse glomerular vascularization was rescued by the CAM derived ECs.

developmental biology↗

MPGEM: A harmonized and transcriptome-complete resource for large-scale reuse of legacy human microarray data

Abstract Background: Legacy microarray datasets provide an extensive record of human transcriptomic biology, but their reuse is constrained by differences in platform design, preprocessing, measurement scale, and gene coverage. Platforms measuring only subsets of genes cannot readily be integrated with higher-coverage platforms, limiting large-scale analysis and computational modeling. Results: We developed Multi-Platform Gene Expression Matrix (MPGEM), a computational framework and resource for harmonizing and completing gene-expression profiles across heterogeneous microarray platforms. MPGEM uses a Reference Quantile Distribution (RQD) and generalized Reference Subset Quantile Distribution (RSQD) framework to transform profiles with different gene coverage onto a common quantitative scale. The MPGEM Engine, a multilayer perceptron, predicts expression of unmeasured genes from genes shared across platforms. Applied to Affymetrix GPL570, GPL571, and GPL96, MPGEM uses GPL570 as a 19,320- gene reference space comprising 12,712 predictor and 6,608 target genes. The resulting resource contains 207,135 human gene-expression profiles across 19,320 genes. Evaluation using masked GPL570 profiles yielded mean sample-wise Pearson and Spearman correlations of 0.944 and 0.939, respectively, and mean gene-wise correlations of 0.830 and 0.825. The lowest-performing 5% of target genes achieved a mean Pearson correlation of 0.683. MPGEM showed comparable or higher predictive performance than baseline mean imputation and K-nearest-neighbor approaches. Conclusions: MPGEM transforms heterogeneous, partially measured legacy microarray profiles into a harmonized, transcriptome-complete representation, facilitating their reuse for large-scale transcriptomic analysis, biomarker discovery, systems biology, and machine learning. The framework, trained models, and expression resource are provided as open-source resources.

bioinformatics↗

An ncBAF-ETS2 Chromatin-Remodelling Axis Drives Vascular Smooth Muscle Cell Osteogenic Reprogramming in Vascular Calcification

Introduction: Vascular calcification is a detrimental ageing-related pathology that is markedly accelerated in metabolic disorders. It is driven by osteogenic differentiation of vascular smooth muscle cells (VSMCs), however epigenetic regulatory pathways activated early in this transition remain poorly defined. Methods: An in vitro calcification model was developed using primary human aortic VSMCs cultured with or without mineral stress. Epigenetic changes were assessed using targeted PCR arrays and CUT&RUN sequencing. Key findings were validated in vivo using single-cell sequencing datasets from human large arteries and spatial transcriptomic analysis in atherosclerotic carotid plaques. Transcriptomic and CUT&RUN analyses identified gene targets altered by epigenetic remodelling, and molecular tools were applied to study effects on metabolism, inflammation, apoptosis, and calcification. Results: During early calcification in response to mineral stress, SWI/SNF chromatin remodelling complexes shift toward ncBAF enrichment in pre-osteogenic VSMCs. ncBAF complexes activated transcriptional programs involved in inflammation, apoptosis, and glycolysis-all hallmarks of calcifying VSMCs. The transcription factor ETS2 was identified as a novel component of ncBAF complexes. Disruption of ncBAF or ETS2 impaired osteogenic differentiation and calcification. Notably, ETS2 expression was regulated by ncBAF, forming a positive feedback loop that reinforced VSMC phenotypic switching. Co-activation of ETS2 and ncBAF and the resulting transcriptional shifts were confirmed in human arterial single-cell datasets, with osteogenic/inflammatory clusters showing NFkB and RUNX2 activation. Spatial transcriptomics further suggested that a macrophage-rich microenvironment may promote the differentiation of smooth muscle cells toward an overt osteogenic/inflammatory phenotype. Immunohistochemistry showed that ETS2 levels correlated with calcification severity in human vessels supporting the potential clinical relevance of ETS2. Conclusions: Our findings identify a novel epigenetic mechanism in vascular calcification, where ncBAF and ETS2 cooperate to drive VSMC phenotypic switching. This ncBAF-ETS2 axis represents a potential therapeutic target to modulate VSMC plasticity and intervene early in the progression of cardiovascular calcification.

cell biology↗

BactoMate: an integrated platform for reproducible bacterial microscopy analysis

Quantitative microscopy of microorganisms increasingly produces large, multidimensional datasets, yet their analysis often depends on fragmented workflows spanning file conversion, segmentation, quality control, fluorescence quantification, tracking, and visualization. Here, we present BactoMate, an open-source, cross-platform graphical user interface that integrates these steps into a unified workflow for microbial image analysis. BactoMate incorporates established segmentation methods and supports both single-file and batch processing. Its modules enable image preprocessing, cell segmentation, morphology-based quality control, fluorescence and foci quantification, single-cell tracking, lineage reconstruction, structured data export, and generation of quality-control and visualization outputs. We demonstrate the applicability of BactoMate across multichannel fluorescence imaging, bacterial swimming assays, microcolony lineage analysis, phage infection assay and a microfluidic time series. All user-configurable parameters are exposed through the interface, are recorded alongside structured outputs and can be loaded for reproducible image analyses across experiments to reduce introduction of bias. By reducing workflow handoffs while preserving parameter control and exportable results, BactoMate enables accessible, reproducible, and scalable quantitative analysis of microbial microscopy data.

microbiology↗

Tracing the regulatory atlas of non-coding RNA in human labour

The early onset of labour increases mortality and developmental risks for a human newborn. Key genes in human labour have been investigated using multiple modalities, but their regulation by non-coding RNA (e.g. lncRNA and miRNA) remains incomplete. This study explores the three-way relationship between labour-associated transcription factors (TFs), miRNA and lncRNA suggested by the competing endogenous RNA (ceRNA) hypothesis, to understand the underlying regulatory framework. Experimentally validated miRNA-lncRNA interactions are modelled using five distinct machine learning (ML) architectures to predict 20469 labour-linked miRNA-lncRNA interactions. Known mRNA-ncRNA interactions from databases were included to construct a tripartite network, and a subset of 9989 labour-linked network motifs containing TFs were isolated and analysed. Gene enrichment of nodes in TF-lncRNA-miRNA network, as well as validation from public myometrial datasets indicate high significance in contractile pathways including immune signalling. Experimentally unconfirmed tripartite network motifs have been found, and we elaborate on their potential regulation in labour using 8 TF-lncRNA-miRNA network motifs. A unified ncRNA-TF regulatory atlas in labour has been synthesized, and a complete summary of the tripartite network motifs can be accessed and visualised using the user-friendly, public database.

bioinformatics↗

RIG-I-like receptors sense replication stress through ribosomal DNA integrity to drive inflammation

Replication stress, a major source of DNA damage and vulnerability in cancer, activates innate immunity through both the cGAS-STING and RIG-I-like receptor (RLR) pathways, yet the mechanism for RLR activation remains poorly understood despite frequent suppression of cGAS-STING in cancer. Here, we show that replication stress induces genomic instability at ribosomal DNA (rDNA), accompanied by aberrant RNA polymerase I-dependent sense and antisense transcription that generates immunostimulatory RNA activating RLR signaling. RLR activation required CDK1 activity but occurred independently of mitosis or micronuclei formation. Cas9-mediated disruption of transcribed rDNA regions was sufficient to induce dysregulated rDNA transcription and RLR activation. Human tumor genomes also showed increased structural variations at rDNA, consistent with rDNA instability. Together, these findings reveal how genomic instability drives inflammation through rDNA-derived immunostimulatory RNA and identify a previously unappreciated consequence of damage-induced transcription.

molecular biology↗

ClonoScreen3D-CRISPRi Uncovers Genetic Modifiers of Radiation Response in Glioblastoma

BackgroundGlioblastoma (GBM) is the most aggressive primary brain tumor in adults. Radioresistance, partly mediated by glioma stem-like cells, represents a major clinical challenge which could be overcome by the identification of the modulators of radioresistance. Existing CRISPR screens in human GBM models have largely used two-dimensional cultures with short-term viability readouts, failing to capture the long-term clonogenic behaviour underlying tumour recurrence after radiotherapy. MethodWe developed ClonoScreen3D-CRISPRi, combining CRISPRi-mediated gene knockdown with three-dimensional clonogenic survival assays. Two GBM cell lines (G7 and GBML20), differing in MGMT promoter methylation status, were engineered to express the KRAB-dCas9 editor. Nine candidate radiosensitivity modifiers, selected through transcriptomic analysis, pharmacological studies, and literature review, were examined in both lines. Target validation was performed using full radiation dose-response assays and a pharmacological inhibitor. ResultsThe majority of candidate genes significantly altered survival fraction following irradiation in both cell lines. Knockdown of NFKB2, RELB, and CDK9 produced the most potent radiosensitization, with sensitizer enhancement ratios of 1.39-1.70 in validation studies -- exceeding those of established radiosensitizers including PARP and ATM inhibitors. Notably, knockdown of these genes induced no significant cytotoxicity in the absence of radiation. Pharmacological validation using an IKK inhibitor confirmed these findings, implicating non-canonical NF-{kappa}B signalling and CDK9-dependent transcriptional elongation as critical adaptive mechanisms in GBM radioresistance. ConclusionsClonoScreen3D-CRISPRi is a scalable, physiologically relevant platform for identifying genetic modifiers of radioresistance. The non-canonical NF-{kappa}B pathway and CDK9 represent promising radiosensitizing targets, and larger screens could enable systematic prioritisation of candidates for clinical translation. Key PointsO_LIClonoScreen3D-CRISPRi combines gene knockdown with 3D clonogenic survival assays C_LIO_LIWe identified NFKB2, RELB, and CDK9 as modifiers of radioresistance in two GBM cell lines C_LIO_LIValidation experiments show ClonoScreen3D-CRISPRi reliably identifies radiosensitizers in GBM C_LI Importance of the studyGlioblastoma (GBM) remains one of the most lethal human cancers, with radioresistance representing a central barrier to improved patient outcomes. While CRISPR-based screens have begun to illuminate genetic drivers of GBM biology, prior approaches using human models have largely relied on two-dimensional culture systems and short-term viability readouts that inadequately model the disease. This study introduces ClonoScreen3D-CRISPRi, a novel platform that integrates CRISPRi-mediated gene knockdown with three-dimensional clonogenic survival assays in patient-derived GBM cells -- more faithfully recapitulating the long-term clonogenic potential that underlies post-radiotherapy recurrence. Using this platform, we identified NFKB2, RELB, and CDK9 as potent genetic modifiers of radioresistance, with sensitizer enhancement ratios exceeding those of established clinical radiosensitizers such as PARP and ATM inhibitors. Pharmacological validation of the non-canonical NF-{kappa}B pathway demonstrates direct translational relevance, providing a rationale for targeting this axis in combination with radiotherapy to improve GBM treatment.

cancer biology↗

The Urinary Tract Commensal Peptoniphilus spp. Encodes a Novel 17β-Hydroxysteroid Dehydrogenase

Microbial steroid metabolism represents an underappreciated extension of the vertebrate endocrine system, with growing evidence that host-associated microbes contribute to the diversity and bioavailability of sex steroids within human tissues. Emerging studies have linked microbial androgen metabolism to urinary microbiome composition and to resistance to androgen deprivation therapy (ADT) in prostate cancer. While microbial pathways capable of converting steroid precursors such as cortisol to androgens, via the steroid-17,20-desmolase pathway, such as DesG-mediated interconversion of androstenedione to testosterone have been reported, the diversity of enzymes mediating downstream androgen interconversion remains incompletely defined. Here, we investigate the androgen-forming capabilities of anaerobic bacteria from the male genitourinary microbiome, focusing on NADPH-dependent 17{beta}-hydroxysteroid dehydrogenases (17{beta}-HSDHs) that catalyze interconversion of androstenedione and testosterone. We isolated androgen-forming bacterial strains from human male urine and identified a previously uncharacterized 17{beta}-HSDH encoded by Peptoniphilus obesi, demonstrated that this enzyme catalyzes the NADPH-dependent reduction of androstenedione to testosterone and the reverse oxidation reaction. Sequence similarity searches further identified a homologous 17{beta}-HSDH in Anaerococcus, which was synthesized and functionally validated, revealing conserved activity despite low sequence identity to the previously characterized urinary tract enzyme DesG. The enzymes were found to have broad substrate specificity for C19 and C18 17keto- and 17{beta}-hydroxysteroids. Together, these findings expand the known diversity of microbial 17{beta}-HSDHs and identify previously unrecognized androgen-forming activities within the genitourinary microbiome. ImportanceMicrobial steroid-transforming pathways may provide a mechanism by which commensal anaerobes contribute to androgen availability in the genitourinary tract. By identifying novel 17{beta}-hydroxysteroid dehydrogenases from Peptoniphilus and Anaerococcus, genera repeatedly associated with prostate cancer, this study provides mechanistic insight into how microbial steroid metabolism may influence hormone-driven disease.

microbiology↗

Epigenetic control of S100A8/A9-driven monocytic inflammation licenses anti-leukemic functionality of immature NK cells during hematopoietic stem cell differentiation.

Inflammation is a key driver of hematopoietic dysfunction in myeloid malignancies, but its role in the context of hypomethylating therapy remains incompletely understood. Although 5-Azacytidine is used posttransplant in high-risk myelodysplastic syndrome (MDS), only 50% of patients show a clinical response. We provide evidence that inherent inflammatory properties of healthy donor CD34+ stem cells exist that are likely to contribute to the "response" seen in MDS patients. These are linked to epigenetic priming of the myeloid niche, resulting in S100A8/A9-driven inflammatory program that promotes functionality of immature NK cells. Using in vitro differentiation systems, multi-omic profiling, and a S100A9-/- mouse model, we find that 5-AzaC modulates inflammatory transcriptional programs through epigenetic rewiring of upstream regulatory elements. Loss of S100A9 disrupts myeloid differentiation, impairs NK cell maturation, and alters key developmental regulators including CEBPB, JUN, and NFIL3. In vivo, 5-AzaC restores these defects and primes NK cells in a time- and context-dependent manner. Re-analysis of the published Australian MDS/CMML cohort shows that "responders" display increased S100A8/A9 expression together with enhanced IFN-{gamma}, IL6-JAK-STAT3, and TNF signaling. These findings suggest that inflammatory myeloid programs may serve as predictive biomarkers and therapeutic targets to enhance NK cell-mediated graft-versus-leukemia activity posttransplant. SummaryO_LIWe provide compelling evidence that inherent properties of healthy donor CD34+ hematopoietic stem cells (SCs) exist that are likely to contribute to the "response" seen upon pre-emptive posttransplant 5-AzaC therapy of patients with high-risk myelodysplastic syndrome (MDS). C_LIO_LIThese properties are linked to a distinct form of epigenetic plasticity at upstream-located transcription factor (TF) binding sites. This may indirectly contribute to acute S100A8/A9-driven inflammation, which is demonstrable in distinct monocyte subsets and, importantly, also in NK cells thereby determining the characteristics of inflammatory monocyte-NK cell crosstalk. C_LIO_LIMice with a targeted deletion of S100A9 fail to upregulate CEBPB / JUN and NFIL3 which results in impaired myeloid priming and dysfunctional NK cell maturation, respectively. C_LIO_LIRe-analysis of the Australian MDS/CMML cohort confirms that MDS patients that "respond" to 5-AzaC exhibit activated IFN-{gamma}, IL6-JAK-STAT3, and TNF-signaling pathways in the context of upregulated S100A8/A9 after six months of treatment. C_LIO_LIOur study indicates that screening of healthy donors SCs for specific inflammatory markers in early developing monocytes could be used as a marker to predict which donor will have the potential of generating a S100A8/A9-driven inflammatory response. This may help identify patients with MDS as well as AML who are likely to benefit from low-dose, short-term 5-AzaC therapy as early as day 7 after transplantation, potentially resulting in increased graft-versus-leukemia (GvL) activity. C_LI

immunology↗

ApeA cleaves genomic RNA to defend against RNA phage infection

To protect themselves against bacteriophage infection, bacteria encode a vast diversity of antiphage defense systems. However, the mechanisms of action of most of these systems have exclusively been studied using phages with DNA genomes as the models, while phages with RNA genomes remain understudied. Here, we investigate how the defense system ApeA confers resistance against RNA phage infection. We show that two ApeA homologs, Ec1ApeA and Ps2ApeA, protect against a variety of single-stranded RNA phages. Focusing on Ec1ApeA, we find that it senses infection through a conserved pocket that likely binds an RNA structure in the phage genome. This activates the HEPN (higher eukaryotes and prokaryotes nucleotide-binding) RNase domain of Ec1ApeA which consequently cleaves the phage genomic RNA to restrict replication. In contrast to many other described defense systems, Ec1ApeA activity directly stops viral replication without inducing cell death, establishing ApeA as a non-abortive defense system that protects against RNA phages. Our results add to the increasingly diverse targets of antiviral HEPN RNases and provide insights into the understudied field of RNA phage defense.

microbiology↗

Direct pathway enrichment prediction from histopathological whole slide images and comparison with gene expression mediated models

Molecular profiling of tumours via RNA sequencing (RNA-seq) enables clinically actionable stratification but remains costly, tissue-intensive, and time-consuming. Recent advances in computational pathology suggest that routine H&E whole-slide images (WSIs) can be utilized to estimate transcriptomic states of cancer cells. Given the WSI-derived predictions of transcriptional signatures are noisy, their use for accurate biological interpretation faces challenges. On the other hand pathway enrichment analysis has been routinely used in describing biologically meaningful cellular states from noisy gene expression data and some studies have evaluated the ability of WSI-predicted gene expression profiles to reconstruct enriched pathways in experiments where the two data modalities were concurrently available. However, it remains unclear if a predictive model that is designed to predict enriched pathways directly from WSI samples would be better than the current approaches to do so by first predicting gene expressions. Here, we develop and evaluate these two complementary approaches for predicting pathway enrichment profiles from WSIs in TCGA Breast Invasive Carcinoma (TCGA-BRCA) by training parallel models which predict pathway enrichment directly from image features and those which rely on predicted gene expression profiles, which is the current state-of-the-art. Our results suggest that under controlled experiments direct prediction of a selected pool of enriched pathways outperforms the models trained on predicting gene expression and then inferring enrichments on predicted gene expression values. These findings will be helpful in prioritizing the goals of predictive modeling of WSI images and improving diagnostic outcomes of cancer patients.

bioinformatics↗

A comprehensive benchmark of publicly available image foundation models for their usability to predict gene expression from whole slide images

Recent advances in large-scale self-supervised learning have led to the emergence of foundation models capable of extracting transferable visual representations from high-dimensional image data. In computational pathology, such models are increasingly used as feature encoders for molecular prediction tasks. However, systematic benchmarking of publicly available image foundation models for transcriptomic prediction from whole-slide images (WSIs) remains limited. Here, we perform a comprehensive evaluation of five state-of-the-art vision foundation models-DINOv2, Phikon, UNI, H-Optimus-0, and MedSigLIP-for gene expression prediction using the TCGA-BRCA cohort. Tile embeddings extracted from each model were aggregated via attention-based multiple instance learning (MIL), followed by multi-target regression to predict RNA-seq expression profiles. Performance was assessed using gene-level Spearman correlation across samples. Histopathology-specific foundation models consistently outperformed general-purpose encoders, with Phikon achieving the strongest overall performance, followed by UNI and H-Optimus-0. These findings demonstrate that domain-aligned pretraining substantially enhances morphology-to-transcriptome inference and provide a principled benchmark for foundation model selection in molecular pathology.

bioinformatics↗