Biology subjects

Savova, V.

Publications and source records attributed to Savova, V..

8 recordsLinked to original sources

bioRxiv2025

Single-cell neural network classifiers reveal that PM21 NK cell expansion is dependent on B cell signaling

In the field drug development ML/AI methods are being applied to improve drug production speed, costs, and reliability. In allogenic NK cell therapy production, one of the biggest challenges is the inherent variability in the donors that provide the starting material for NK cell expansion. In this study we performed PM21-mediated NK cell expansion on 26 donors, and in parallel performed single-cell transcriptomics on the same donor sample prior to expansion. Canonical differential expression analysis and cell state abundance did not highlight any significant difference between donors with high and low NK cell expansion yield. Instead, training neural networks classifiers for high-yield donors enabled identifying several highly predictive models with perfect cross-validation recall. Further investigation of the most predictive models unveiled a previously unknown role for B cell in supportive NK cell expansion. Overall, this study represents a blueprint for combining deep phenotyping and machine learning methods to unveil novel biology and improve the quality and speed of delivery of cell therapeutics to patients.

systems biology↗

bioRxiv2025

An Agentic AI Framework for Ingestion and Standardization of Single-Cell RNA-seq Data Analysis

The proliferation of publicly available single-cell RNA sequencing (scRNA-seq) data has created significant opportunities in biomedical research. However, the reuse of these resources is constrained by a series of preparatory steps, including metadata extraction from primary literature, retrieval of datasets from corresponding repositories, and the subsequent manual execution of standardized downstream analysis. These tasks often require manual scripting and rely on fragmented workflows, limiting accessibility and increasing turnaround time. To address these challenges, we designed a two-component system consisting of an artificial intelligence (AI) agent coordinating an automated analysis pipeline. CellAtria (Agentic Triage of Regulated single-cell data Ingestion and Analysis) is an agentic AI framework that enables dialogue-driven, document-to-analysis automation through a chatbot interface. Built on a graph-based, multi-actor architecture, CellAtria integrates a large language model (LLM) with tool-execution capabilities to orchestrate the full lifecycle of data reuse. To support downstream analysis, CellAtria incorporates CellExpress, a co-developed pipeline that applies state-of-the-art scRNA-seq processing steps to transform raw count matrices into analysis-ready single-cell profiles. Thus, CellAtria provides computational skill-agnostic and time-efficient access to standardized single-cell data ingestion and analysis.

bioinformatics↗

bioRxiv2024

Immune disease dialogue of chemokine-based cell communications as revealed by single-cell RNA sequencing meta-analysis

Immune-mediated diseases are characterized by aberrant immune responses, posing significant challenges to global health. In both inflammatory and autoimmune diseases, dysregulated immune reactions mediated by tissue-residing immune and non-immune cells precipitate chronic inflammation and tissue damage that is amplified by peripheral immune cell extravasation into the tissue. Chemokine receptors are pivotal in orchestrating immune cell migration, yet deciphering the signaling code across cell types, diseases and tissues remains an open challenge. To delineate disease-specific cell-cell communications involved in immune cell migration, we conducted a meta-analysis of publicly available single-cell RNA sequencing (scRNA-seq) data across diverse immune diseases and tissues. Our comprehensive analysis spanned multiple immune disorders affecting major organs: atopic dermatitis and psoriasis (skin), chronic obstructive pulmonary disease and idiopathic pulmonary fibrosis (lung), ulcerative colitis (colon), IgA nephropathy and lupus nephritis (kidney). By interrogating ligand-receptor (L-R) interactions, alterations in cell proportions, and differential gene expression, we unveiled intricate disease- specific and common immune cell chemoattraction and extravasation patterns. Our findings delineate disease-specific L-R networks and shed light on shared immune responses across tissues and diseases. Insights gleaned from this analysis hold promise for the development of targeted therapeutics aimed at modulating immune cell migration to mitigate inflammation and tissue damage. This nuanced understanding of immune cell dynamics at the single-cell resolution opens avenues for precision medicine in immune disease management.

immunology↗

bioRxiv2024

An immunocompetent human kidney on-a-chip model to study renal inflammation and immune-mediated injury

Kidney damage and dysfunction is an emerging health issue worldwide resulting in high morbidity and mortality rates. Numerous renal diseases are recognized to be driven by the immune system. Despite this recognition, the development of targeted therapies has been challenging as knowledge of the underlying mechanism and complex interactions remains insufficient. Recent advancements in the field offer promising avenues for exploring the interplay between renal cells and immune cells and their role in the development of renal inflammation and diseases. This study describes the establishment of a human immunocompetent 3D in vitro co-culture model of the proximal tubule in a high-throughput microfluidic platform that can be used to study renal functionality and inflammatory processes. The model incorporated RPTEC in the top compartment and HUVECs in the bottom compartment cultured under flow and in direct contact with a collagen-I ECM gel resulting in the formation of polarized tubular structures. As an immune component, human primary monocytes of different donors were added to the lumen of the endothelium. Renal inflammation was successfully induced using complement activated serum (CAS) as evident by epithelial morphological changes, increased expression of adhesion molecules, release of pro-inflammatory cytokines, and reduced epithelial viability. Realtime migratory behavior of monocytes showed increased extravasation and migration towards the ECM and Renal compartment upon exposure to CAS with donor-to-donor differences observed. Finally, immune modulatory compounds showed efficacious inhibition of monocyte migration under inflammatory conditions in the microfluidic co-culture model. A successful co-culture model was established and can be applied to study renal functionality in health and disease but also for drug screening due to the compatibility of the platform with automation and relatively high throughput. Overall, the described proximal tubule model has high potential to fill the gap that currently exists to study renal inflammation preclinically.

immunology↗

bioRxiv2023

Comparative Analysis of Association Networks Using Single-Cell RNA Sequencing Data Reveals Perturbation-Relevant Gene Signatures

Single-cell RNA sequencing (scRNA-seq) data has elevated our understanding of systemic perturbations to organismal physiology at the individual cell level. However, despite the rich information content of scRNA-seq data, the relevance of genes to a perturbation is still commonly assessed through differential expression analysis. This approach provides a one-dimensional perspective of the transcriptomic landscape, risking the oversight of tightly controlled genes characterized by modest changes in expression but with profound downstream effects. We present GENIX (Gene Expression Network Importance eXamination), a novel platform for constructing gene association networks, equipped with an innovative network-based comparative model to uncover condition-relevant genes. To demonstrate the effectiveness of GENIX, we analyze influenza vaccine-induced immune responses in peripheral blood mononuclear cells (PBMCs) collected from recovered COVID-19 patients, shedding light on the mechanistic underpinnings of gender differences. Our methodology offers a promising avenue to identify genes relevant to perturbation responses in biological systems, expanding the scope of response signature discovery beyond differential gene expression analysis. HIGHLIGHTSO_LIConventional methods used to identify perturbation-relevant genes in scRNA-seq data rely on differential expression analysis, susceptible to overlooking essential genes. C_LIO_LIGENIX leverages cell-type-specific inferred gene association networks to identify condition-relevant genes and gene programs, irrespective of their specific expression alterations. C_LIO_LIGENIX provides insight into the gene-regulatory response to the influenza vaccine in naive and recovered COVID-19 patients, expanding on previously observed gender-specific differences. C_LI GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=115 SRC="FIGDIR/small/556872v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@1837d3eorg.highwire.dtl.DTLVardef@1937860org.highwire.dtl.DTLVardef@c40114org.highwire.dtl.DTLVardef@22d3b9_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2022

DiSiR: a software framework to identify ligand-receptor interactions at subunit level from single-cell RNA-sequencing data

Most of cell-cell interactions and crosstalks are mediated by ligand-receptor interactions. The advent of single-cell RNA-sequencing (scRNA-seq) techniques has enabled characterizing tissue heterogeneity at single-cell level. Over the past recent years, several methods have been developed to study ligand-receptor interactions at cell type level using scRNA-seq data. However, there is still no easy way to query the activity of a specific user-defined signaling pathway in a targeted way or map the interactions of the same subunit with different ligands as part of different receptor complexes. Here, we present DiSiR, a fast and easy-to-use permutation-based software framework to investigate how individual cells are interacting with each other by analyzing signaling pathways of multi-subunit ligand-activated receptors from scRNA-seq data, not only for available curated databases of ligand-receptor interactions, but also for interactions that are not listed in these databases. We show that, when utilized to infer melanoma disease map on a gold-standard dataset, DiSiR outperforms other well-known permutation-based methods, e.g., CellPhoneDB and ICELLNET. To demonstrate DiSiRs utility in exploring data and generating biologically relevant hypotheses, we apply it to COVID lung and rheumatoid arthritis (RA) synovium scRNA-seq data and highlight potential differences between inflammatory pathways at cell type level for control vs. disease samples.

cell biology↗

bioRxiv2021

Cell type classification and discovery across diseases, technologies and tissues reveals conserved gene signatures and enables standardized single-cell readouts

Autoimmune diseases are a major cause of mortality1,2. Current treatments often yield severe insult to host tissue. It is hypothesized that improved "precision medicine" therapies will target pathogenic cells selectively and thus reduce or eliminate severe side effects, and potentially induce robust immune tolerance3. However, it remains challenging to systematically identify which cellular phenotypes are present in cellular ensembles. Here, we present a novel machine learning approach, Signac, which uses neural networks trained with flow-sorted gene expression data to classify cellular phenotypes in single cell RNA-sequencing data. We demonstrate that Signac accurately classified single cell RNA-sequencing data across diseases, technologies, species and tissues. Then we applied Signac to identify known and novel immune-relevant precision medicine candidate drug targets (n = 12) in rheumatoid arthritis. A full release of this workflow can be found at our GitHub repository (https://github.com/mathewchamberlain/Signac).

immunology↗

bioRxiv2019

An optimized workflow for single-cell transcriptomics and repertoire profiling of purified lymphocytes from clinical samples.

Establishing clinically relevant single-cell (SC) transcriptomic workflows from cryopreserved tissue is essential to move this emerging immune monitoring technology from the bench to the bedside. Improper sample preparation leads to detrimental cascades, resulting in loss of precious time, money and finally compromised data. There is an urgent need to establish protocols specifically designed to overcome the inevitable variations in sample quality resulting from uncontrollable factors in a clinical setting. Here, we explore sample preparation techniques relevant to a range of clinically relevant scenarios, where SC gene expression and repertoire analysis are applied to a cryopreserved sample derived from a small amount of blood, with unknown or partially known preservation history. We compare a total of ten cell-counting, viability-improvement, and lymphocyte-enrichment methods to highlight a number of unexpected findings. Trypan blue-based automated counters, typically recommended for single-cell sample quantitation, consistently overestimate viability. Advanced sample clean-up procedures significantly impact total cell yield, while only modestly increasing viability. Finally, while pre-enrichment of B cells from whole peripheral blood mononuclear cells (PBMCs) results in the most reliable BCR repertoire data, comparable T-cell enrichment strategies distort the ratio of CD4+ and CD8+ cells. Furthermore, we provide high-resolution analysis of gene expression and clonotype repertoire of different B cell subtypes. Together these observations provide both qualitative and quantitative sample preparation guidelines that increase the chances of obtaining high-quality single-cell transcriptomic and repertoire data from human PBMCs in a variety of clinical settings.

immunology↗