5 minutes with Dr John Szilagyi on PanHunter

Dr John Szilagyi

John is a research scientist in the non-clinical investigative toxicology department at Bristol Meyers Squibb (BMS). His area of focus is placental toxicology. He has a broad expertise in drug transport, enzyme and cell-based assays. John obtained his PhD in toxicology from Rutgers University. He joined BMS in June 2020 after a postdoctoral research stay in the toxicology and environmental medicine department of the University of North Carolina at Chapel Hill.

In what project do you use PanHunter?

John: My active collaboration with Evotec involves high-content screening and bioinformatics investigations of differentiating pluripotent stem cells treated with protein-degrading teratogens, such as thalidomide. The overall goal of this work is to better understand the mechanisms behind stem cell differentiation in embryonic limb development and identify the sensitive pathways that can disrupt that process. In doing so, we hope to better inform drug discovery and development to improve the development strategy for safer and more effective medicines.

How does PanHunter leverage research outcomes at BMS?

John: PanHunter^TM is a great platform to ergonomically see large bioinformatics datasets at multiple levels, allowing a researcher to easily build the context behind the observed biological changes. Notably in our current investigations, PanHunter allows the direct identification of high priority pathways during stem cell differentiation that can lead to teratogenicity. These data can then be integrated into drug discovery to prioritize safer treatments for patients.

What is the impact of PanHunter in your daily business?

John: I spend 80% of my working time in the lab developing assays to assess underlying mechanisms of unexpected findings or problems encountered by other colleagues. For this purpose, we often use omics analyses. PanHunter speeds up and eases the analysis of these multi-omics datasets. More particular, PanHunter comes with a broad spectrum of tools to help me to really understand the mechanisms behind these unexpected findings.

What are the main points where PanHunter improved or accelerated your analysis?

John: The experimental setup of our multi-omics study is quite complex. Hence, there are many different angles to look at and analyze the data. I joked that I would need a post-doc to look at the data for three years to fully understand the findings. The interactivity of PanHunter makes it very easy for me dive into, sorting and pulling out data. My ultimate goal is to understand the mode of- action of the compounds. It is very easy to set up differential analysis in PanHunter. To analyze the outcomes of these differential analyses I started with Euler and Venn diagrams to understand differences and similarities in the mechanisms of the compounds. Another very useful tool is the network visualization that is based on BioGRID. This helped me a lot to identify the proteins that have physical interactions with the target protein.

What could you do faster/better because you had PanHunter at hand?

John: Definitively the identification of proteins that have physical interactions with the target protein. As well as the detection of pathways that are affected by the compounds. This helped me a lot to understand the mechanism behind the compounds. A thing that I could do much faster is the hierarchical clustering of abundances of proteins/transcripts based on predefined features lists. The predefined feature lists originate either from public databases like GO or from earlier research that I did. Thank you for the interview.