Biology subjects

Ghaffari, S.

Publications and source records attributed to Ghaffari, S..

9 recordsLinked to original sources

bioRxiv2026

Machine Learning-Guided Discovery of Bacterial-Selective Membrane-Active Compounds Reveals Mechanistic Bias in Antibiotic Training Datasets

The rise of antibiotic resistance necessitates the discovery of antibacterial compounds with novel mechanisms of action (MoAs). Recent machine learning approaches have shown promise in antibacterial compound discovery, but often identify derivatives of known antibiotic classes rather than mechanistically novel compounds. Previous approaches applied Tanimoto similarity filters at the end of screening pipelines, but this method has substantial drawbacks: Tanimoto similarity can be misleading in chemical space, and post-hoc filtering does not influence what activity models learn to prioritize. Here, we present a machine learning pipeline that addresses chemical novelty upfront by employing an XGBoost-based MoA classifier to explicitly prioritize compounds predicted to have mechanisms distinct from known antibiotic classes, combined with graph neural networks for antibacterial activity and toxicity prediction. Applied to the Zinc20 database, our approach successfully identified non-toxic antibacterial compounds structurally distinct from known antibiotics. Notably, the majority of these hits exhibited membrane-targeting activity with selectivity for bacterial cells over mammalian cells, suggesting potential for next-generation membrane-active antibiotics. However, we did not identify compounds with novel protein targets. Systematic analysis revealed that this limitation stems from mechanistic bias in training data rather than model architecture. Specifically, our activity model learned to preferentially score compounds similar to specific groups in the training data, thus overrepresenting certain MoA classes including membrane-active compounds. Even substantial model architecture and training data enhancements did not overcome this constraint. Our findings demonstrate that the primary bottleneck for discovering mechanistically novel antibiotics is the scarcity of diverse, mechanistically-annotated training data. This work provides both a methodological framework for mechanism-aware screening and critical insights into data requirements for genuinely novel antibiotic discovery.

bioinformatics↗

bioRxiv2025

Dissociable Spatial and Feature Tuning of Gamma and Alpha Rhythms in Human Visual Cortex

Visual stimulation in humans reliably induces narrowband gamma (40-80 Hz) increases and alpha (8-12 Hz) suppression in EEG, but the relationship between these rhythms is not fixed. Using high-density EEG, we mapped gamma and alpha responses across a broad set of retinotopic, orientation, and motion conditions. Gamma was strongly retinotopically tuned, with linear summation of subfield responses accurately predicting full-field responses. In contrast, alpha showed little spatial tuning, substantial responses even in the absence of visual input (anticipatory suppression), and subadditive summation consistent with divisive normalization. Across most retinotopic configurations, higher gamma coincided with stronger alpha suppression, yet systematic dissociations emerged, with full-field and foveal-centered gratings evoking stronger than expected gamma relative to alpha suppression. Orientation tuning was robust for gamma but weak for alpha, with oblique gratings producing high gamma yet weak alpha suppression, reversing the usual inverse relationship. These patterns indicate that EEG gamma power primarily reflects large-scale synchronous activity that matches the global LFP observed in macaque V1, whose spatial and feature tuning properties are independent of both single-neuron selectivity and feedback-driven alpha dynamics. The results establish a mechanistic dissociation between gamma and alpha rhythms, highlighting distinct circuit origins and tuning principles for these canonical visual responses.

neuroscience↗

bioRxiv2023

Using mechanistic models and machine learning to design single-color multiplexed Nascent Chain Tracking experiments

mRNA translation is the ubiquitous cellular process of reading messenger-RNA strands into functional proteins. Over the past decade, large strides in microscopy techniques have allowed observation of mRNA translation at a single-molecule resolution for self-consistent time-series measurements in live cells. Dubbed Nascent chain tracking (NCT), these methods have explored many temporal dynamics in mRNA translation uncaptured by other experimental methods such as ribosomal profiling, smFISH, pSILAC, BONCAT, or FUNCAT-PLA. However, NCT is currently restricted to the observation of one or two mRNA species at a time due to limits in the number of resolvable fluorescent tags. In this work, we propose a hybrid computational pipeline, where detailed mechanistic simulations produce realistic NCT videos, and machine learning is used to assess potential experimental designs for their ability to resolve multiple mRNA species using a single fluorescent color for all species. Through simulation, we show that with careful application, this hybrid design strategy could in principle be used to extend the number of mRNA species that could be watched simultaneously within the same cell. We present a simulated example NCT experiment with seven different mRNA species within the same simulated cell and use our ML labeling to identify these spots with 90% accuracy using only two distinct fluorescent tags. The proposed extension to the NCT color palette should allow experimentalists to access a plethora of new experimental design possibilities, especially for cell signalling applications requiring simultaneous study of multiple mRNAs.

systems biology↗

bioRxiv2022

A Robust Bayesian Approach to Bulk Gene Expression Deconvolution withNoisy Reference Signatures

BackgroundDifferential gene expression in bulk transcriptomics data can reflect change of transcript abundance within a cell type and/or change in the proportion of cell types within the sample. Expression deconvolution methods can help differentiate these scenarios and enable more accurate inference of gene regulation by estimating the contributions of individual cell types to bulk transcriptomic profiles. However, the accuracy of these methods is sensitive to technical and biological differences between bulk profiles and the cell type-signatures required by them as references. ResultsWe present BEDwARS, a Bayesian deconvolution method specifically designed to address differences between reference signatures and the unknown true signatures underlying bulk transcriptomic profiles. Through extensive benchmarking utilizing eight different datasets derived from pancreas and brain, we demonstrate that BEDwARS outperforms leading in-class methods for estimating cell type proportions and signatures. Furthermore, we systematically show that BEDwARS is more robust to noisy reference signatures than all compared methods. Finally, we apply BEDwARS to newly generated RNA-seq and scRNA-seq data on over 100 induced pluripotent stem cell-derived neural organoids to study mechanisms underlying a rare pediatric condition (Dihydropyridine Dehydrogenase deficiency), identifying the possible involvement of ciliopathy and impaired translational control in the etiology of the disorder. ConclusionWe propose a new approach to bulk gene expression deconvolution which estimates the cell type proportions and cell type signatures simultaneously and is robust to commonly seen mismatches between reference and true cell type signatures. Application of our method lead to novel findings about mechanisms of a rare pediatric condition.

bioinformatics↗

bioRxiv2022

The PREGCARE study: Personalized recurrence risk assessment following the birth of a child with a pathogenic de novo mutation

Next-generation sequencing has led to a dramatic improvement in molecular diagnoses of serious pediatric disorders caused by apparently de novo mutations (DNMs); by contrast, clinicians ability to counsel the parents about the risk of recurrence in a future child has lagged behind. Owing to the possibility that one of the parents could be mosaic in their germline, a recurrence risk of 1-2% is frequently quoted, but for any specific couple, this figure is usually incorrect. We present a systematic approach to providing individualized recurrence risk stratification, by combining deep-sequencing of multiple tissues in the mother-father-child trio with haplotyping to determine the parental origin of the DNM. In the first 58 couples analysed (total of 59 DNMs in 49 different genes), the risk for 35 (59%) DNMs was decreased below 0.1% but for 6 (10%) couples it was increased owing to parental mosaicism - that could be quantified in semen (recurrence risks of 5.6-12.1%) for the paternal cases. Deep-sequencing of the DNM efficiently identifies couples at greatest risk for recurrence and may qualify them for additional reproductive technologies. Haplotyping can further reassure many other couples that their recurrence risk is very low, but its implementation is more technically challenging and will require better understanding of how couples respond to information that reduces their risks.

genetics↗

bioRxiv2022

Elevated p21 (CDKN1a) mediates β-thalassemia erythroid apoptosis but its loss does not improve β-thalassemic erythropoiesis

{beta}-thalassemias are common hemoglobinopathies due to mutations in the {beta}-globin gene that lead to hemolytic anemias. Premature death of {beta}-thalassemic erythroid precursors results in ineffective erythroid maturation, increased production of erythropoietin (Epo), expansion of erythroid progenitor compartment, extramedullary erythropoiesis and splenomegaly. However, the molecular mechanism of erythroid apoptosis in {beta}-thalassemia is not well understood. Using a mouse model of {beta}-thalassemia (Hbbth3/+), we show that dysregulated expression of Foxo3 transcription factor and its upstream pro-apoptotic regulator TP53 is implicated in {beta}-thalassemia erythroid apoptosis. In Foxo3-/- /Hbbth3/+ mice, erythroid apoptosis is significantly reduced while erythroid cell maturation, red blood cell and hemoglobin production are substantially improved. However, persistence of elevated reticulocytes and splenomegaly suggests that ineffective erythropoiesis is not resolved in Foxo3-/-/Hbbth3/+. We next focused on cell cycle inhibitor Cdkn1a (p21) and show that p21 that is a target of both Foxo3 and TP53 is markedly upregulated in both mouse and patients-derived {beta}-thalassemic erythroid precursors. To address the contribution of p21 to {beta}-thalassemia pathophysiology, we generated p21-/- /Hbbth3/+ mice. Double mutant p21/Hbbth3/+ mice exhibited embryonic lethality with only a fraction of mice surviving to weaning. Notably, studies in adult mice showed apoptosis and circulating Epo were greatly reduced in erythroid compartments of surviving p21-/- /Hbbth3/+ relative to Hbbth3/+ mice, while ineffective erythroid cell maturation, extramedullary erythropoiesis and splenomegaly were not modified. These combined results indicate that while lack of Foxo3 reduces apoptosis and improves anemia, diminished p21-mediated apoptosis is insufficient to improve red blood cell production in Hbbth3/+ mice. They also suggest that a molecular network constituted by p21, FOXO3 and TP53, control erythroid cell survival and differentiation in {beta}-thalassemia. Overall, these studies provide a new framework for investigating ineffective erythropoiesis in {beta}-thalassemia. Key PointsO_LIElevated p21 mediates {beta}-thalassemia erythroid cell apoptosis C_LIO_LILoss of Foxo3 or p21 reduces {beta}-thalassemia erythroid cell apoptosis but does not improve {beta}-thalassemia ineffective erythropoiesis C_LIO_LIA network of Foxo3, p21 and TP53 controls {beta}-thalassemia erythroid apoptosis C_LIO_LIApoptosis may be uncoupled from ineffective erythropoiesis in {beta}-thalassemia C_LI

developmental biology↗

bioRxiv2020

Simulating flow induced migration in vascular remodelling

Shear stress induces directed endothelial cell (EC) migration in blood vessels leading to vessel diameter increase and induction of vascular maturation. Other factors, such as EC elongation and interaction between ECs and non-vascular areas are also important. Computational models have previously been used to study collective cell migration. These models can be used to predict EC migration and its effect on vascular remodelling during embryogenesis. We combined live time-lapse imaging of the remodelling vasculature of the quail embryo yolk sac with flow quantification using a combination of micro-Particle Image Velocimetry and computational fluid dynamics. We then used the flow and remodelling data to inform a model of EC migration during remodelling. To obtain the relation between shear stress and velocity in vitro for EC cells, we developed a flow chamber to assess how confluent sheets of ECs migrate in response to shear stress. Using these data as an input, we developed a multiphase, self-propelled particles (SPP) model where individual agents are driven to migrate based on the level of shear stress while maintaining appropriate spatial relationship to nearby agents. These agents elongate, interact with each other, and with avascular agents at each time-step of the model. We compared predicted vascular shape to real vascular shape after 4 hours from our time-lapse movies and performed sensitivity analysis on the various model parameters. Our model shows that shear stress has the largest effect on the remodelling process. Importantly, however, elongation played an especially important part in remodelling. This model provides a powerful tool to study the input of different biological processes on remodelling. Author summaryShear stress is known to play a leading role in endothelial cell (EC) migration and hence, vascular remodelling. Vascular remodelling is, however, more complicated than only EC migration. To achieve a better understanding of this process, we developed a computational model in which, shear stress mediated EC migration has the leading role and other factors, such as avascular regions and EC elongation, are also accounted for. We have tested this model for different vessel shapes during remodelling and could study the role that each of these factors play in remodelling. This model gives us the possibility of addition of other factors such as biochemical signals and angiogenesis which will help us in the study of vascular remodelling in both development and disease.

developmental biology↗

bioRxiv2019

Oregano essential oil vapour prevents Plasmopora viticola infection in grapevine (Vitis Vinifera) by triggering autoimmune metabolic pathways

The reduction of synthetic fungicides in agriculture a major challenge in maintaining sustainable production, protecting the environment and consumers health. Downy mildew caused by the oomycete Plasmopora viticola is the major pathogen in viticulture worldwide and responsible for up to 60% of pesticide treatments. Alternatives to reduce fungicides are thus utterly needed to ensure sustainable vineyard-ecosystems, consumer health and public acceptance. Essential oils (EOs) are amongst the most promising natural plant protection alternatives and have shown their antibacterial, antiviral and antifungal properties on several agricultural crops. However, the efficiency of EOs highly depends on timing, application method and the molecular interactions between the host, the pathogen and EO. Despite proven EO efficiency, the underlying processes are still not understood and remain a black box. The objectives of the present study were: a) to evaluate whether a continuous fumigation of a particular EO can control downy mildew in order to circumvent the drawbacks of direct application, b) to decipher molecular mechanisms that could be triggered in the host and the pathogen by EO application and c) to try to differentiate whether essential oils directly repress the oomycete or act as plant resistance primers.\n\nA custom-made climatic chamber was used for a continuous fumigation of potted vines with different EOs during long-term experiments. The grapevine (Vitis vinifera) cv Chasselas was chosen in reason of its high susceptibility to Plasmopara viticola. Grapevine cuttings were infected with P. viticola. and subsequently exposed to continuous fumigation of different EOs at different concentrations, during 2 application time spans (24 hours and 10 days). Experiments were stopped when infection symptoms were clearly observed on the leaves of the control plants. Plant physiology (photosynthesis and growth rate parameters) were recorded and leaves were sampled at different time points for subsequent RNA extraction and transcriptomics analysis. Strikingly, the Oregano vulgare essential oil vapour treatment during 24h post-infection proved to be sufficient to reduce downy mildew development by 95%. Total RNA was extracted from leaves of 24h and 10d treatments and used for whole transcriptome shotgun sequencing (RNA-seq). Sequenced reads were then mapped onto the V. vinifera and P. viticola genomes. Less than 1% of reads could be mapped onto the P. viticola genome from treated samples, whereas up to 30 % reads from the controls mapped onto the P. viticola genome, thereby confirming the visual observation of P. viticola absence in the treated plants. On average, 80 % of reads could be mapped onto the V. vinifera genome for differential expression analysis, which yielded 4800 modulated genes. Transcriptomic data clearly showed that the treatment triggered the plants innate immune system with genes involved in salicylic, jasmonic acid and ethylene synthesis and signaling, activating Pathogenenesis-Related-proteins as well as phytoalexin synthesis.\n\nThese results elucidate EO-host-pathogen interactions for the first time and indicate that the antifungal efficiency of EO is mainly due to the triggering of resistance pathways inside the host plants. This is of major importance for the production and research on biopesticides, plant stimulation products and for resistance-breeding strategies\n\nAuthor SummaryThe reduction of synthetic plant protection products is a major concern of modern agriculture. The oomycete Plasmopora viticola which causes downy mildew in grapevine is amongst the most important grapevine pests and responsible for the dispersion of huge amounts of pesticides in vineyards. Among the evaluated alternatives to reduce or replace synthetic pesticides, plant volatile compounds could represent a sustainable solution. Some plant essential oils (EOs) have already shown antifungal capacities. However, their application is often difficult in terms of the right timing of treatment, degradation, bad rainfastness, mixability and phytotoxicity.\n\nThe aim of the present work was to investigate whether the vapour phase, applied by a continuous fumigation of different EOs, might inhibit the development of downy mildew on grapevine, and in case of proven efficiency, to study the induced transcriptomic changes by RNA-sequencing in an attempt to elucidate the underlaying molecular interactions. Our results showed that the vapour phases of O. vulgare and T. vulgaris were highly efficient against the pathogen. The study of differentially expressed genes indicated that the EO vapour triggered the main mechanisms of the plants innate immune system such as PTI (Pattern-Triggered Immunity) and ETI (Effector Triggered immunity).\n\nFor the first time these results highlight the effects of EOs vapour on plant genes expression, which is very valuable information for the development of new natural plant protection products, as well as for breeding disease resistant cultivars.

plant biology↗

bioRxiv2018

Fyn Is Involved In Erythropoietin Signaling Pathway And Interfaces Oxidation To Regulate Erythropoiesis

Erythropoiesis is a complex multistep process responsible of the production of circulating mature erythrocytes and involved the production of reactive oxygen species (ROS) during erythroid differentiation. Here, we document that Fyn, a Src-family-kinase, participates in erythropoietin (EPO) signaling pathway, by the reducing extent of Tyr-phosphorylation of EPO-R and by decreasing STAT5 activity. The importance of Fyn in EPO cascade is also supported by the increased sensitivity of Fyn-/- mice to stress erythropoiesis. Fyn -/- mouse erythroblasts adapt to the induced stress by the activation of the redox-related-transcription-factor Nrf2. However, the absence of the Nrf2 physiologic repressor Fyn resulted in the persistent activation of Nrf2 and accumulation of non-functional proteins. This is paralleled by ROS induced over-activation of Jak2-Akt-mTOR pathway and repression of autophagy and perturbation of lysosomal-clearance during Fyn -/- reticulocyte maturation. Treatment with Rapamycin, a mTOR inhibitor and autophagy activator, ameliorates Fyn-/- mouse baseline erythropoiesis and restored the erythropoietic response to phenylhydrazine. Taken together these findings have enabled to identify the novel multimodal action of Fyn in the developmental program of erythropoiesis.

cell biology↗