Biology subjects

Rezaei, M.

Publications and source records attributed to Rezaei, M..

5 recordsLinked to original sources

bioRxiv2025

Small extracellular vesicles reflect senescence progression in human bone marrow-derived mesenchymal stem cells during hollow fiber bioreactor culture

Prolonged three-dimensional culture exposes stem cells to sustained microenvironmental and mechanical stresses that can promote aging- and senescence-associated phenotypic alterations. This study examined how long-term expansion of human bone marrow-derived mesenchymal stem cells (BMSCs) in a hollow fiber bioreactor (HFB) influences cellular senescence and the molecular composition of secreted small extracellular vesicles (sEVs). During extended HFB culture, BMSCs exhibited progressive morphological flattening and cytoskeletal disorganization, accompanied by increased senescence-associated {beta}-galactosidase activity and immunophenotypic remodeling characterized by reduced fluorescence intensity and spatial redistribution of canonical MSC markers, consistent with a stress-adapted, early senescence-associated cellular state. In parallel, sEVs were collected longitudinally over 40 days and characterized by nanoparticle tracking analysis, immunoblotting, and quantitative proteomics. While vesicle size, marker expression, and yield remained stable throughout culture, proteomic profiling revealed pronounced, phase-dependent remodeling of sEV cargo, including coordinated alterations in oxidative stress-related processes, lysosomal and extracellular matrix-associated pathways, and relative depletion of cytoskeletal and translational components. Notably, these vesicular signatures closely mirrored senescence-associated changes observed at the cellular level. The strong correspondence between cellular phenotypes and sEV proteomic profiles establishes vesicle analysis as a convergent and noninvasive readout of BMSC aging, enabling sensitive monitoring of senescence progression while reducing reliance on parallel, labor-intensive cellular assays. Collectively, these findings indicate that prolonged HFB culture promotes a controlled, stress-associated senescence program in BMSCs and position sEV proteomic profiling as a robust approach for assessing stem cell aging dynamics during long-term three-dimensional bioreactor culture.

molecular biology↗

bioRxiv2025

Predicting Kinase-Substrate Phosphorylation Site Using Autoregressive Transformer

Accurately predicting kinase-specific phosphorylation sites remains difficult due to the diversity of kinases and the context-dependent nature of substrate recognition. Importantly, aberrant kinase overactivation is a hallmark of many cancers including colorectal, gastric, liver, and breast tumors where dysregulated kinase signaling promotes malignant transformation, tumor progression, and therapy resistance. This underscores the clinical importance of understanding kinase-substrate relationships and precisely mapping phosphorylation events. In this paper, we introduce two complementary sequence-based architectures that operate directly on full-length substrate and kinase sequences. Stage 1 extends a task-agnostic prediction method, named Prot2Token, to jointly support three tasks: kinase-group classification from substrate sequences alone, kinase-substrate interaction prediction, and kinase-specific phosphorylation-site prediction while incorporating a self-supervised decoder pretraining task that predicts amino-acid positions from encoder embeddings. This pretraining substantially strengthens site prediction. Stage 2 specializes the architecture for phosphorylation-site prediction by replacing causal decoding of Prot2Token with a bidirectional one, yielding further gains. On standard benchmarks, the specialized model consistently outperforms widely used baselines. Beyond in-distribution evaluation, across both in-distribution and zero-shot settings of understudied dark kinases, we show the sign of zero-shot kinase-specific phosphorylation-site prediction capability. Together, these results indicate that jointly modeling substrate and kinase sequences provides a straight-forward, scalable approach to state-of-the-art, zero-shot-capable phosphorylationsite prediction.

bioinformatics↗

bioRxiv2025

Using Autoregressive-Transformer Model for Protein-Ligand Binding Site Prediction

AO_SCPLOWBSTRACTC_SCPLOWAccurate prediction of protein-ligand binding sites is critical for understanding molecular interactions and advancing drug discovery. Existing computational approaches often suffer from limited generality, restricting their applicability to a small subset of ligands, while data scarcity further impairs performance, particularly for underrepresented ligand types. To address these challenges, we introduce a unified model that integrates a protein language model with an autoregressive transformer for protein-ligand binding site prediction. By framing the task as a language modeling problem and incorporating task-specific tokens, our method achieves broad ligand coverage while relying solely on protein sequence input. We systematically analyze ligand-specific task token embeddings, demonstrating that they capture meaningful biochemical properties through clustering and correlation analyses. Furthermore, our multi-task learning strategy enables effective knowledge transfer across ligands, significantly improving predictions for those with limited training data. Experimental evaluations on 41 ligands highlight the models superior generalization and applicability compared to existing methods. This work establishes a scalable generative AI framework for binding site prediction, laying the foundation for future extensions incorporating structural information and richer ligand representations. The code, model, and datasets are available at this link.

bioinformatics↗

bioRxiv2024

Pre-trained molecular representations enable antimicrobial discovery

The rise in antimicrobial resistance poses a worldwide threat, reducing the efficacy of common antibiotics. Determining the antimicrobial activity of new chemical compounds through experimental methods is still a time-consuming and costly endeavor. Compound-centric deep learning models hold the promise to speed up this search and prioritization process. Here, we introduce a lightweight computational strategy for antimicrobial discovery that builds on MolE(Molecular representation through redundancy reduced Embedding), a deep learning framework that leverages unlabeled chemical structures to learn task-independent molecular representations. By combining MolE representation learning with experimentally validated compound-bacteria activity data, we design a general predictive model that enables assessing compounds with respect to their antimicrobial potential. The model correctly identified recent growth-inhibitory compounds that are structurally distinct from current antibiotics and discovered de novo three human-targeted drugs as Staphylococcus aureus growth inhibitors which we experimentally confirmed. Our framework offers a viable cost-effective strategy to accelerate antibiotics discovery.

microbiology↗

bioRxiv2023

Mechanoreceptive Aβ primary afferents discriminate naturalistic social touch inputs at a functionally relevant time scale

Interpersonal touch is an important part of our social and emotional interactions. How these physical, skin-to-skin touch expressions are processed in the peripheral nervous system is not well understood. From single-unit microneurography recordings in humans, we evaluated the capacity of six subtypes of cutaneous afferents to differentiate perceptually distinct social touch expressions. By leveraging conventional statistical analyses and classification analyses using convolutional neural networks and support vector machines, we found that single units of multiple A{beta} subtypes, especially slowly adapting type II (SA-II) and fast adapting hair follicle afferents (HFA), can reliably differentiate the skin contact of those expressions at accuracies similar to those perceptually. Rapidly adapting field (Field) afferents exhibit lower accuracies, whereas C-tactile (CT), fast adapting Pacinian corpuscles (FA-II), and muscle spindle (MS) afferents can barely differentiate the expressions, despite responding to the stimuli. We then identified the most informative firing patterns of SA-II and HFA afferents spike trains, which indicate that an average duration of 3-4 s of firing provides sufficient discriminative information. Those two subtypes also exhibit robust tolerance to shifts in spike-timing of up to 10 ms. A greater shift in spike-timing, however, drastically compromises an afferents discrimination capacity, and can change a firing patterns envelope to resemble that of another expression. Altogether, the findings indicate that SA-II and HFA afferents differentiate the skin contact of social touch at time scales relevant for such interactions, which is 1-2 orders of magnitude longer than those relevant for discriminating non-social touch inputs.

neuroscience↗