Biology subjects

Yousefi, S.

Publications and source records attributed to Yousefi, S..

11 recordsLinked to original sources

bioRxiv2025

Mitochondrial Dysfunction Rewires Macrophage Metabolism, Driving Pro-inflammatory Priming and Immune System Remodeling

Macrophage activation is tightly coupled to cellular metabolism: classically activated pro-inflammatory (M1) macrophages rely on glycolysis and a disrupted tricarboxylic acid cycle, whereas alternatively activated (M2) macrophages depend on oxidative phosphorylation (OXPHOS) and fatty acid oxidation. Although mitochondria are central to this metabolic plasticity, it remains unclear whether mitochondrial dysfunction itself can dictate macrophage polarization. Using macrophage-specific OPA1 knockout mice, we investigated how mitochondrial dysfunction influences macrophage metabolism and immune homeostasis. Loss of OPA1 caused severe impairment of OXPHOS, reduced mitochondrial membrane potential, and a compensatory glycolytic shift, driving M0 and M2 macrophages toward an M1-like bioenergetic state. Integrative metabolomic and transcriptomic analyses revealed strong priming of OPA1-deficient macrophages towards classical activation, including accumulation of M1-associated metabolites (lactate, succinate, itaconate) and upregulation of NF-{kappa}B-driven and other inflammatory gene programs, resulting in increased secretion of IL-6 and TNF even in the absence of stimulation. Functionally, this metabolic shift primed non-activated and M2 macrophages toward partial M1 polarization with enhanced bactericidal capacity, while simultaneously suppressing M2-associated processes such as proliferation, efferocytosis, and expression of Arg1, CD206, and RELM. In vivo, OPA1{Delta}M mice displayed reduced peritoneal macrophage abundance, impaired self-renewal after IL-4 complex stimulation, and compensatory monocyte recruitment. The remaining macrophages exhibited increased MHCII and reduced RELM expression, consistent with partial M1 skewing and loss of alternative activation. These local alterations were mirrored systemically: blood profiling revealed enhanced T-cell activation and sex-specific remodeling of immune composition during inflammation and aging. Collectively, these findings demonstrate that mitochondrial dysfunction serves as a cell-intrinsic cue that primes macrophages toward a glycolytic, pro-inflammatory phenotype while constraining their M2 properties and proliferative capacities. This dual metabolic and functional rewiring highlights mitochondrial integrity as a pivotal determinant of macrophage immunometabolic identity and reveals how its disruption can reshape both local and systemic immune homeostasis.

immunology↗

bioRxiv2024

Creating a biomedical knowledge base by addressing GPT inaccurate responses and benchmarking context

We created GNQA, a generative pre-trained transformer (GPT) knowledge base driven by a performant retrieval augmented generation (RAG) with a focus on aging, dementia, Alzheimers and diabetes. We uploaded a corpus of three thousand peer reviewed publications on these topics into the RAG. To address concerns about inaccurate responses and GPT hallucinations, we implemented a context provenance tracking mechanism that enables researchers to validate responses against the original material and to get references to the original papers. To assess the effectiveness of contextual information we collected evaluations and feedback from both domain expert users and citizen scientists on the relevance of GPT responses. A key innovation of our study is automated evaluation by way of a RAG assessment system (RAGAS). RAGAS combines human expert assessment with AI-driven evaluation to measure the effectiveness of RAG systems. When evaluating the responses to their questions, human respondents give a "thumbs-up" 76% of the time. Meanwhile, RAGAS scores 90% on answer relevance on questions posed by experts. And when GPT-generates questions, RAGAS scores 74% on answer relevance. With RAGAS we created a benchmark that can be used to continuously assess the performance of our knowledge base. Full GNQA functionality is embedded in the free GeneNetwork.org web service, an open-source system containing over 25 years of experimental data on model organisms and human. The code developed for this study is published under a free and open-source software license at https://git.genenetwork.org/gn-ai/tree/README.md.

scientific communication and education↗

bioRxiv2023

Human enteric glia diversity in health and disease: new avenues for the treatment of Hirschsprung disease

Hirschsprung disease (HSCR) is caused by an absence of the enteric nervous system (ENS), which is crucial for intestinal function. The ENS is composed of enteric neurons and glia, and is mostly derived from migrating vagal neural crest cells. Trunk-derived Schwann cells also play a significant role in postnatal maintenance of the ENS. However, the diversity of the ENS in health and disease remains largely unknown. Here, we performed single cell RNA sequencing on pediatric controls and HSCR individuals, and identified two major classes of enteric glia, being canonical and Schwann-like enteric glia. We show that the latter are the main contributors of enteric glia heterogeneity after birth and importantly, that they are preserved in aganglionic segments of HSCR individuals. In a zebrafish model of HSCR, which also shows preservation of Schwann-like enteric glia, enteric neurogenesis could be stimulated, demonstrating a potential novel therapy for HSCR.

pathology↗

bioRxiv2021

Comprehensive multi-omics integration identifies differentially active enhancers during human brain development with clinical relevance

BackgroundNon-coding regulatory elements (NCREs), such as enhancers, play a crucial role in gene regulation and genetic aberrations in NCREs can lead to human disease, including brain disorders. The human brain is complex and can be affected by numerous disorders; many of these are caused by genetic changes, but a multitude remain currently unexplained. Understanding NCREs acting during brain development has the potential to shed light on previously unrecognised genetic causes of human brain disease. Despite immense community-wide efforts to understand the role of the non-coding genome and NCREs, annotating functional NCREs remains challenging. ResultsHere we performed an integrative computational analysis of virtually all currently available epigenome data sets related to human fetal brain. Our in-depth analysis unravels 39,709 differentially active enhancers (DAEs) that show dynamic epigenomic rearrangement during early stages of human brain development, indicating likely biological function. Many of these DAEs are linked to clinically relevant genes, and functional validation of selected DAEs in cell models and zebrafish confirms their role in gene regulation. Compared to enhancers without dynamic epigenomic rearrangement, these regions are subjected to higher sequence constraints in humans, have distinct sequence characteristics and are bound by a distinct transcription factor landscape. DAEs are enriched for GWAS loci for brain related traits and for genetic variation found in individuals with neurodevelopmental disorders, including autism. ConclusionOur compendium of high-confidence enhancers will assist in deciphering the mechanism behind developmental genetics of the human brain and will be relevant to uncover missing heritability in human genetic brain disorders.

genomics↗

bioRxiv2020

Computational approaches towards reducing contamination in single-cell RNA-seq data

Single cell RNA sequencing has enabled quantification of single cells and identification of different cell types and subtypes as well as cell functions in different tissues. Single cell RNA sequence analyses assume acquired RNAs correspond to cells, however, RNAs from contamination within the input data are also captured by these assays. The sequencing of background contamination as well as unwanted cells making their way to the final assay Potentially confound the correct biological interpretation of single cell transcriptomic data. Here we demonstrate two approaches to deal with background contamination as well as profiling of unwanted cells in the assays. We use three real-life datasets of whole-cell capture and nucleotide single-cell captures generated by Fluidigm and 10x technologies and show that these methods reduce the effect of contamination, strengthen clustering of cells and improves biological interpretation.

bioinformatics↗

bioRxiv2019

Predicting glaucoma prior to its onset using deep learning

PurposeTo assess the accuracy of deep learning models to predict glaucoma development from fundus photographs several years prior to disease onset.\n\nDesignA deep learning model for prediction of glaucomatous optic neuropathy or visual field abnormality from color fundus photographs.\n\nParticipantsWe retrospectively included 66,721 fundus photographs from 3,272 eyes of 1,636 subjects to develop deep leaning models.\n\nMethodFundus photographs and visual fields were carefully examined by two independent readers from the optic disc and visual field reading centers of the ocular hypertension treatment study (OHTS). When an abnormality was detected by the readers, subject was recalled for re-testing to confirm the abnormality and further confirmation by an endpoint committee. Using OHTS data, deep learning models were trained and tested using 85% of the fundus photographs and further validated (re-tested) on the remaining (held-out) 15% of the fundus photographs.\n\nMain Outcome MeasuresAccuracy and area under the receiver-operating characteristic curve (AUC).\n\nResultsThe AUC of the deep learning model in predicting glaucoma development 4-7 years prior to disease onset was 0.77 (95% confidence interval 0.75, 0.79). The accuracy of the model in predicting glaucoma development about 1-3 years prior to disease onset was 0.88 (0.86, 0.91). The accuracy of the model in detecting glaucoma after onset was 0.95 (0.94, 0.96).\n\nConclusionsDeep learning models can predict glaucoma development prior to disease onset with reasonable accuracy. Eyes with visual field abnormality but not glaucomatous optic neuropathy had a higher tendency to be missed by deep learning algorithms.

bioengineering↗

bioRxiv2019

Loss of UGP2 in brain leads to a severe epileptic encephalopathy, emphasizing that bi-allelic isoform specific start-loss mutations of essential genes can cause genetic diseases

Developmental and/or epileptic encephalopathies (DEEs) are a group of devastating genetic disorders, resulting in early onset, therapy resistant seizures and developmental delay. Here we report on 12 individuals from 10 families presenting with a severe form of intractable epilepsy, severe developmental delay, progressive microcephaly and visual disturbance. Whole exome sequencing identified a recurrent, homozygous variant (chr2:64083454A>G) in the essential UDP-glucose pyrophosphorylase (UGP2) gene in all probands. This rare variant results in a tolerable Met12Val missense change of the longer UGP2 protein isoform but causes a disruption of the start codon of the shorter isoform. We show that the absence of the shorter isoform leads to a reduction of functional UGP2 enzyme in brain cell types, leading to altered glycogen metabolism, upregulated unfolded protein response and premature neuronal differentiation, as modelled during pluripotent stem cell differentiation in vitro. In contrast, the complete lack of all UGP2 isoforms leads to differentiation defects in multiple lineages in human cells. Reduced expression of Ugp2a/Ugp2b in vivo in zebrafish mimics visual disturbance and mutant animals show a behavioral phenotype. Our study identifies a recurrent start codon mutation in UGP2 as a cause of a novel autosomal recessive DEE. Importantly, it also shows that isoform specific start-loss mutations causing expression loss of a tissue relevant isoform of an essential protein can cause a genetic disease, even when an organism-wide protein absence is incompatible with life. We provide additional examples where a similar disease mechanism applies.

genetics↗

bioRxiv2019

Transfer Learning From Nucleus Detection To Classification In Histopathology Images

Despite significant recent success, modern computer vision techniques such as Convolutional Neural Networks (CNNs) are expensive to apply to cell-level prediction problems in histopathology images due to difficulties in providing cell-level supervision. This work explores the transferability of features learned by an object detection CNN (Faster R-CNN) to nucleus classification in histopathology images. We detect nuclei in these images using class-agnostic models trained on small annotated patches, and use the CNN representations of detected nuclei to cluster and classify them. We show that with a small training dataset, the proposed pipeline can achieve superior nucleus detection and classification performance, and generalizes well to unseen stain types.

bioinformatics↗

bioRxiv2018

Gold Nanoprisms as Optical Coherence Tomography Contrast Agents in the Second Near Infrared Window for Enhanced Angiography in Live Animals

Optical coherence tomography angiography (OCTA) is an important tool for investigating vascular networks and microcirculation in living tissue. Traditional OCTA detects blood vessels via intravascular dynamic scattering signals derived from the movements of red blood cells (RBCs). However, the low hematocrit and long latency between RBCs in capillaries makes these OCTA signals discontinuous, leading to incomplete mapping of the vascular networks. OCTA imaging of microvascular circulation is particularly challenging in tumors due to the abnormally slow blood flow in angiogenic tumor vessels and strong attenuation of light by tumor tissue. Here we demonstrate in vivo that gold nanoprisms (GNPRs) can be used as OCT contrast agents working in the second near infrared window, significantly enhancing the dynamic scattering signals in microvessels and improving the sensitivity of OCTA in skin tissue and melanoma tumors in live mice. This is the first demonstration that nanoparticle-based OCT contrast agent work in vivo in the second near infrared window, which allows deeper imaging depth by OCT. With GNPRs as contrast agents, the post-injection OCT angiograms showed 41% and 59% more microvasculature than pre-injection angiograms in healthy mouse skin and melanoma tumors, respectively. By enabling better characterization of microvascular circulation in vivo, GNPR-enhanced OCTA could lead to better understanding of vascular functions during pathological conditions, more accurate measurements of therapeutic response, and improved patient prognoses.

bioengineering↗

bioRxiv2017

Predicting cancer outcomes from histology and genomics using convolutional networks

Cancer histology reflects underlying molecular processes and disease progression, and contains rich phenotypic information that is predictive of patient outcomes. In this study, we demonstrate a computational approach for learning patient outcomes from digital pathology images using deep learning to combine the power of adaptive machine learning algorithms with traditional survival models. We illustrate how this approach can integrate information from both histology images and genomic biomarkers to predict time-to-event patient outcomes, and demonstrate performance surpassing the current clinical paradigm for predicting the survival of patients diagnosed with glioma. We also provide techniques to visualize the tissue patterns learned by these deep learning survival models, and establish a framework for addressing intratumoral heterogeneity and training data deficits.

bioinformatics↗

bioRxiv2017

Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models

Translating the vast data generated by genomic platforms into accurate predictions of clinical outcomes is a fundamental challenge in genomic medicine. Many prediction methods face limitations in learning from the high-dimensional profiles generated by these platforms, and rely on experts to hand-select a small number of features for training prediction models. In this paper, we demonstrate how deep learning and Bayesian optimization methods that have been remarkably successful in general high-dimensional prediction tasks can be adapted to the problem of predicting cancer outcomes. We perform an extensive comparison of Bayesian optimized deep survival models and other state of the art machine learning methods for survival analysis, and describe a framework for interpreting deep survival models using a risk backpropagation technique. Finally, we illustrate that deep survival models can successfully transfer information across diseases to improve prognostic accuracy. We provide an open-source software implementation of this framework called SurvivalNet that enables automatic training, evaluation and interpretation of deep survival models.

bioinformatics↗