SYSTEMS BIOLOGY APPLICATION OF PERTURBATION GENE EXPRESSION PROFILES
Barsi Szilvia
Molekuláris Orvostudományok
Dr. Várnai Péter
SE Élettani Intézet könyvtára
2026-01-29 13:00:00
Celluláris és molekuláris élettan
Dr. Hunyady László
Dr. Hunyady László, Dr. Szalai Bence
Dr. Fichó Erzsébet
Dr. Dobson László Imre
Dr. Hegedűs Tamás
Dr. Bögel Gábor
Dr. Ligeti Balázs
This work demonstrates applicability of perturbation gene expression signatures as a tool for drug repurposing and receptor activity inference, highlighting their value in uncovering mechanisms of action and guiding hypothesis generation.
First, I leveraged SARS-CoV-2 infection-induced gene expression signatures and compared them with drug treatment-induced perturbation profiles to gain insights into antiviral drug action. In contrast to the classical signature-reversal approach, we found that effective antiviral drugs often mimic adaptive host responses, activating pathways like NFkB and JAK-STAT. Several of these drugs also activated SREBF1/2, key regulators of lipid metabolism. Experimental validation using fluorescent cholesterol sensors confirmed that these drugs reduce plasma membrane cholesterol and that this depletion contributes to their antiviral effect. These findings refine our understanding of signature-based drug repurposing in viral contexts and highlight cholesterol modulation as a key antiviral mechanism.
I demonstrated the systematic use of chemical and genomic perturbation signatures. I developed RIDDEN (Receptor actIvity Data Driven inferENce), a computational tool that infers receptor activity from transcriptomic profiles of 229 receptors. Unlike co-expression-based methods, RIDDEN leverages the downstream transcriptional response of receptor perturbation to provide insights into cell-cell communication. Benchmarking on in vitro and in vivo datasets showed that RIDDEN accurately reflects the transcriptional responses of the cytokine receptor modulation and performs comparably to, or better than, existing models in this context. When applied to a cancer immunotherapy dataset, RIDDEN identified receptor activities associated with treatment response, revealing biologically meaningful signals that gene expression is not able to capture.
These applications demonstrate the value of perturbation gene expression signatures and how they can serve as a bridge between high-dimensional data and interpretable biological insight, guiding hypothesis generation in drug discovery and advancing our understanding of complex cellular responses.
