Mostafa Rezapour

Mostafa Rezapour


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
MRMostafa Rezapour
Mostafa Rezapour
TSThomas D. Shupe
Thomas D. Shupe
DODavid A. Ornelles
David A. Ornelles
Sean MurphySean Murphy
Sean Murphy
AAAnthony Atala
Anthony Atala
Dengue virus (DENV) and chikungunya virus (CHIKV) co-circulate in many regions and present with overlapping clinical features, which complicate accurate diagnosis and disease management. This study develops an integrative transcriptomic framework to identify robust host gene signatures that distinguish between dengue, chikungunya, and healthy states. Publicly available RNA sequencing (RNA-seq) datasets derived from human blood samples were analyzed using a cross-validation design to ensure robustness and prevent information leakage. Differential expression analysis was performed independently within each dataset using the Generalized Linear Models with Quasi-Likelihood F-tests and Magnitude–Altitude Scoring (GLMQL-MAS) framework, followed by Cross-Magnitude–Altitude Scoring (Cross-MAS) integration to identify shared and virus-specific gene signatures. A strict consensus approach across folds was applied to derive reproducible gene sets. These signatures were used for dimensionality reduction and multinomial logistic regression to evaluate classification performance. A small subset of selected genes showed strong discriminative performance within the cross-validation framework, with test balanced accuracy reaching 0.97, which improved upon models using all genes. Biologically, both infections exhibited a shared antiviral response characterized by interferon signaling and innate immune activation. However, distinct virus-specific patterns were identified. Dengue infection was associated with cell-cycle and DNA replication pathways, while chikungunya infection showed stronger enrichment of inflammatory and immune signaling pathways, including NF-kappaB and Toll-like receptor signaling. Overall, this study provides a cross-validation-based framework for integrative transcriptomic analysis and identifies compact, reproducible host-response signatures with strong discriminative signals in the analyzed cohorts. These signatures require validation in larger independent cohorts before any clinical or diagnostic application.

17 April 2026 in Frontiers in Genetics
Verified
Journal Article
Open access
MRMostafa Rezapour
Mostafa Rezapour
SMSean V. Murphy
Sean V. Murphy
DODavid A. Ornelles
David A. Ornelles
TSThomas D. Shupe
Thomas D. Shupe
SWStephen J. Walker
Stephen J. Walker
AAAnthony Atala
Anthony Atala
Introduction: Ebola virus (EBOV) infection triggers intense host transcriptional responses that overlap extensively with those induced by other viral and bacterial pathogens. This overlap complicates the identification of EBOV-specific gene expression signatures and limits diagnostic specificity. Defining transcriptional markers that distinguish EBOV from other infections is essential for improving molecular diagnostics and advancing understanding of EBOV-specific host responses. Methods: We developed a multi-step filtering framework using blood-derived RNA-Seq data from nonhuman primates and human cohorts organized into independent training and test sets. In the training cohort, differential expression analysis was performed using an edgeR-based GLMQL-MAS approach to identify EBOV-associated genes. Candidates were filtered against non-EBOV comparator datasets, including mpox virus, influenza, bacterial pneumonia, acute HIV-1 infection, and multiple SARS-CoV-2 variants, to remove broadly shared host-response genes. Genes included in the NanoString nCounter® Host Response Panel were additionally excluded. The resulting EBOV-specific signature was evaluated in independent EBOV and non-EBOV test cohorts using principal component analysis and logistic regression. Functional enrichment was assessed using KEGG pathways. Results: Initial analysis identified numerous interferon-stimulated genes that were similarly upregulated across infections. After cross-infection filtering and NanoString exclusion, 281 EBOV-specific genes were identified. Optimization within the training cohort yielded a top-50 gene set that clearly separated EBOV from Non-EBOV samples. In the independent test cohort, classification performance improved substantially, with the F1 score increasing from 37.5% when all genes were used to 95.0% after applying the top-50 gene set. Enrichment analysis of the top-50 EBOV-specific genes revealed significant association with vascular, coagulation, secretory, and metabolic pathways. ADAMTS1 showed consistent upregulation in EBOV while remaining downregulated or inactive in comparator infections. Discussion: Structured cross-pathogen filtering enables identification of EBOV-specific transcriptional features beyond shared antiviral responses. The validated gene signature generalizes across independent cohorts and highlights biologically distinct pathways, which supports its potential utility for host-based diagnostic development.

8 September 2026 in Frontiers in Genetics
Verified
Journal Article
Open access
PDF file
MRMostafa Rezapour
Mostafa Rezapour
DODavid A. Ornelles
David A. Ornelles
SWStephen J. Walker
Stephen J. Walker
CBColin E. Bishop
Colin E. Bishop
SMSean V. Murphy
Sean V. Murphy
AAAnthony Atala
Anthony Atala
Introduction: Respiratory syncytial virus A (RSVA) and respiratory syncytial virus B (RSVB) may induce distinct airway host responses. We compared them with influenza A virus (IAV) and human metapneumovirus (hMPV) in airway organ tissue equivalents (OTEs) and developed the Relaxed Magnitude Altitude Score with Hub-Centric Network Analysis framework (RMAS-HCNA). Methods: OTEs were infected with RSVA-GFP5 or RSVB-GFP3 at low multiplicity of infection and profiled at 24 and 72 h using the NanoString nCounter Host Response Panel. Previous IAV and hMPV profiles were integrated. Analyses included source-matched differential expression, cross-virus ranking, network-based module selection, nested cross-validation, and Gene Ontology Biological Process enrichment. Results: RSVA showed the strongest temporal expansion and developed a late response centered on intracellular transport, localization, and organelle organization. RSVB remained the most restricted response, shifting from early inflammatory and signaling-associated activity to a later interferon-stimulated profile. IAV maintained a broad antiviral response, whereas hMPV combined limited induction with extensive late downregulation of stress, catabolic, metabolic, and transport-related processes. A shared antiviral component was present across infections, while downregulated responses were largely virus dependent. Under five-fold outer and four-fold inner nested cross-validation, the selected genes achieved a mean outer-test Macro F1 of 0.859 and balanced accuracy of 0.888, compared with 0.604 and 0.616 for the all-gene comparator. Discussion: The four viruses shared an antiviral foundation but followed distinct temporal and functional trajectories. The framework identified compact, biologically representative panels that generalized better than the complete measured gene set, although independent genome-wide and in vivo validation remains necessary.

7 January 2026 in arXiv
Verified
Preprint
Open access
PDF file
MRMostafa Rezapour
Mostafa Rezapour
Differential gene expression (DGE) analysis is foundational to transcriptomic research, yet tool selection can substantially influence results. This study presents a comprehensive comparison of two widely used DGE tools, edgeR and DESeq2, using real and semi-simulated bulk RNA-Seq datasets spanning viral, bacterial, and fibrotic conditions. We evaluated tool performance across three key dimensions: (1) sensitivity to sample size and robustness to outliers; (2) classification performance of uniquely identified gene sets within the discovery dataset; and (3) generalizability of tool-specific gene sets across independent studies. First, both tools showed similar responses to simulated outliers, with Jaccard similarity between the DEG sets from perturbed and original (unperturbed) data decreasing as more outliers were added. Second, classification models trained on tool-specific genes showed that edgeR achieved higher F1 scores in 9 of 13 contrasts and more frequently reached perfect or near-perfect precision. Dolan-More performance profiles further indicated that edgeR maintained performance closer to optimal across a greater proportion of datasets. Third, in cross-study validation using four independent SARS-CoV-2 datasets, gene sets uniquely identified by edgeR yielded higher AUC, precision, and recall in classifying samples from held-out datasets. This pattern was consistent across folds, with some test cases achieving perfect separation using edgeR-specific genes. In contrast, DESeq2-specific genes showed lower and more variable performance across studies. Overall, our findings highlight that while DESeq2 may identify more DEGs even under stringent significance conditions, edgeR yields more robust and generalizable gene sets for downstream classification and cross-study replication, which underscores key trade-offs in tool selection for transcriptomic analyses.