Biology subjects

Yousefi, B.

Publications and source records attributed to Yousefi, B..

9 recordsLinked to original sources

bioRxiv2025

BioChemAIgent: An AI-driven Protein Modeling and Docking Framework for Structure-Based Drug Discovery

Recent advances in AI have substantially accelerated in silico drug discovery, yet most computational approaches remain task-specific and require expert knowledge. We present BioChemAIgent, an agentic framework that orchestrates state-of-the-art AI models and established computational chemistry tools to support end-to-end small-molecule analysis, protein modeling, molecular docking, and interaction analysis through a unified interface. BioChemAIgent emphasizes transparent reasoning, reproducible workflows, and community-oriented extensibility for structural biology and drug discovery applications.

bioinformatics↗

bioRxiv2025

SCALE: Unsupervised Multi-Scale Domain Identification in Spatial Omics Data

Single-cell spatial transcriptomics enables precise mapping of cellular states and functional domains within their native tissue environment. These functional domains often exist at multiple spatial scales, with larger domains encompassing smaller ones, reflecting the hierarchical organization of biological systems. However, the identification of these functional domain hierarchies has been hardly explored due to the lack of appropriate computational methods. In this work, we present SCALE, an unsupervised algorithm for multi-scale domain identification in spatial transcriptomics data. SCALE combines neural graph representation learning with an entropy-based search algorithm to detect functional domains at different scales. It reaches state-of-the-art performance in single- and multi-scale domain detection on simulated and murine brain Xenium and MERFISH data, as well as patient-derived kidney tissue, highlighting its robustness and scalability across diverse tissue types and platforms. SCALEs ease of use makes it a powerful aid for advancing our understanding of tissue organization and function in health and disease.

bioinformatics↗

bioRxiv2024

Spatio-temporal interaction of immune and renal cells determines glomerular crescent formation in autoimmune kidney disease

Rapidly progressive glomerulonephritis (RPGN) is the most aggressive group of autoimmune kidney disease with the worst prognosis. Anti-neutrophil cytoplasmic antibody (ANCA) associated vasculitis, anti-glomerular basement membrane (anti-GBM) and lupus nephritis are the most common causes of RPGN and are characterized by the formation of glomerular crescents and infiltration of leukocytes that eventually lead to glomerulosclerosis and kidney failure. In this work, we used high-resolution spatial transcriptomics of 32 ANCA, 19 lupus nephritis, 6 anti-GBM, and 6 control patients to understand how intercellular signaling between immune and renal tissue cells leads to renal inflammation and glomerular injury. Using 3,218,210 immune and kidney cells, we observed that the biological pathways involved in the sequence of glomerular crescent formation are similar across the diseases. While innate immune cells infiltrated the glomerular compartment relatively early, later increases in adaptive immune cells were largely restricted to the periglomerular regions. These changes in immune cells temporally correlated with increases in glomerular parietal epithelial (PEC) and fibrotic mesangial cells, suggesting disease-relevant functional signaling between these immune and renal cells. Cell communication analysis revealed early disease PDGF signaling from epithelial and mesangial cells to PECs, causing their activation and proliferation. At later stages, TGF-{beta} signaling from macrophages, T cells, epithelial cells, and mesangial cells to PECs triggered the expression of extracellular matrix components resulting in glomerulosclerosis. Our results highlight a spatio-temporally conserved progression into glomerular crescents and sclerosis for ANCA, lupus nephritis, and anti-GBM disease, which is driven by consecutive PDGF and TGF-{beta} signaling to PECs.

immunology↗

bioRxiv2024

PCA-based spatial domain identification with state-of-the-art performance

The identification of biologically meaningful domains is a central step in the analysis of spatial transcriptomic data. Following Occams razor, we show that a simple PCA-based algorithm for spatial domain identification rivals the performance of ten competing state-of-the-art methods across six single-cell spatial transcriptomic datasets. Our reductionist approach, NichePCA, provides researchers with intuitive domain interpretation and excels in execution speed, robustness, and scalability.

bioinformatics↗

bioRxiv2024

Consensus Clustering for Robust Bioinformatics Analysis

Clustering plays an important role in a multitude of bioinformatics applications, including protein function prediction, population genetics, and gene expression analysis. The results of most clustering algorithms are sensitive to variations of the input data, the clustering algorithm and its parameters, and individual datasets. Consensus clustering (CC) is an extension to clustering algorithms that aims to construct a robust result from those clustering features that are invariant under the above sources of variation. As part of CC, stability scores can provide an idea of the degree of reliability of the resulting clustering. This review structures the CC approaches in the literature into three principal types, introduces and illustrates the concept of stability scores, and illustrates the use of CC in applications to simulated and real-world gene expression datasets. Open-source R implementations for each of these CC algorithms are available in the GitHub repository: https://github.com/behnam-yousefi/ConsensusClustering

bioinformatics↗

bioRxiv2023

QPPLab: A generally applicable software package for detecting, analyzing, and visualizing large-scale quasiperiodic spatiotemporal patterns (QPPs) of brain activity

Quasi-periodic patterns (QPPs) are prominent spatiotemporal brain dynamics observed in functional neuroimaging data, reflecting the alternation of high and low activity across brain regions and their propagation along cortical gradients. QPPs have been linked to neural processes such as attention, arousal fluctuations, and cognitive function. Despite their significance, existing QPP analysis tools are limited by study-specific parameters and complex workflows. To address these challenges, we present QPPLab, an open-source MATLAB-based toolbox for detecting, analyzing, and visualizing QPPs from fMRI time series. QPPLab integrates correlation-based iterative algorithms, supports customizable parameter settings, and features automated workflows to simplify analysis. Processing times vary depending on dataset size and the selected mode, with the fast detection mode completing analyses that can be 4-6 times faster than the robust detection mode. Results include spatiotemporal templates of QPPs, sliding correlation time courses, and functional connectivity maps. By reducing manual parameter adjustments and providing user-friendly tools, QPPLab enables researchers to efficiently study QPPs across diverse datasets and species, advancing our understanding of intrinsic brain dynamics. Metadata O_TBL View this table: org.highwire.dtl.DTLVardef@1f4f44dorg.highwire.dtl.DTLVardef@d31544org.highwire.dtl.DTLVardef@19384corg.highwire.dtl.DTLVardef@1bf2360org.highwire.dtl.DTLVardef@1ab9b80_HPS_FORMAT_FIGEXP M_TBL C_TBL

bioinformatics↗

bioRxiv2023

Capturing the dynamics of microbiomes using individual-specific networks

BackgroundLongitudinal analysis of multivariate individual-specific microbiome profiles over time or across conditions remains a daunting task. The vast majority of statistical tools and methods available to study the microbiota are based upon cross-sectional data. Over the past few years, several attempts have been made to model the dynamics of bacterial species over time or across conditions. However, the field needs novel views on how to incorporate individual-specific microbial associations in temporal analyses when the focus lies on microbial interactions. ResultsHere, we propose a novel data analysis framework, called MNDA, to uncover taxon neighbourhood dynamics that combines representation learning and individual-specific microbiome co-occurrence networks. We show that tracking local neighbourhood dynamics in microbiome interaction or co-occurrence networks can yield complementary information to standard approaches that only use microbial abundances or pairwise microbial interactions. We use cohort data on infants for whom microbiome data was available at 6 and 9 months after birth, as well as information on mode of delivery and diet changes over time. In particular, MNDA-based prediction models outperform traditional prediction models based on individual-specific abundances, and enable the detection of microbes whose neighbourhood dynamics are informative of clinical variables. We further show that similarity analyses of individuals based on microbial neighbourhood dynamics can be used to find subpopulations of individuals with potential relevance to clinical practice. The annotated source code for the MNDA framework can be downloaded from: https://github.com/H2020TranSYS/microbiome_dynamics ConclusionsMNDA extracts information from matched microbiome profiles and opens new avenues to personalized prediction or stratified medicine with temporal microbiome data.

microbiology↗

bioRxiv2020

Propagating patterns of intrinsic activity along macroscale gradients coordinate functional connections across the whole brain

The intrinsic activity of the human brain, observed with resting-state fMRI (rsfMRI) and functional connectivity, exhibits macroscale spatial organization such as resting-state networks (RSNs) and functional connectivity gradients (FCGs). Dynamic analysis techniques have shown that the time-averaged maps captured by functional connectivity are mere summaries of time-varying patterns with distinct spatial and temporal characteristics. A better understanding of these patterns might provide insight into aspects of the brains intrinsic activity that cannot be inferred by functional connectivity, RSNs or FCGs. Here, we describe three spatiotemporal patterns of coordinated activity across the whole brain obtained by averaging similar ~20-second-long segments of rsfMRI timeseries. In each of these patterns, activity propagates along a particular macroscale FCG, simultaneously across the cortical sheet and in most other brain regions. In some areas, like the thalamus, the propagation suggests previously-undescribed FCGs. The coordinated activity across areas is consistent with known tract-based connections, and nuanced differences in the timing of peak activity between brain regions point to plausible driving mechanisms. The magnitude of correlation within and particularly between RSNs is remarkably diminished when these patterns are regressed from the rsfMRI timeseries, a quantitative demonstration of their significant role in functional connectivity. Taken together, our results suggest that a few recurring patterns of propagating intrinsic activity along macroscale gradients give rise to and coordinate functional connections across the whole brain.

neuroscience↗

bioRxiv2017

Quasi-Periodic Patterns of Intrinsic Brain Activity in Individuals and their Relationship to Global Signal

Quasiperiodic patterns (QPPs) as reported by Majeed et al., 2011 are prominent features of the brains intrinsic activity that involve important large-scale networks (default mode, DMN; task positive, TPN) and are likely to be major contributors to widely used measures of functional connectivity. We examined the variability of these patterns in 470 individuals from the Human Connectome Project resting state functional MRI dataset. The QPPs from individuals can be coarsely categorized into two types: one where strong anti-correlation between the DMN and TPN is present, and another where most areas are strongly correlated. QPP type could be predicted by an individuals global signal, with lower global signal corresponding to QPPs with strong anti-correlation. After regression of global signal, all QPPs showed strong anti-correlation between DMN and TPN. QPP occurrence and type was similar between a subgroup of individuals with extremely low motion (or even high motion) and the rest of the sample, which shows that motion is not a major contributor to the QPPs. After regression of estimates of slow respiratory and cardiac induced signal fluctuations, more QPPs showed strong anti-correlation between DMN and TPN, an indication that while physiological noise influences the QPP type, it is not the primary source of the QPP itself. QPPs were more similar for the same subjects scanned on different days than for different subjects. These results provide the first assessment of the variability in individual QPPs and their relationship to physiological parameters.

neuroscience↗