What’s Next for Large-Scale Proteomics? With Chris Whelan, Janssen

·23 min·2 clips

A machine learning model with just five proteins predicted multiple myeloma a decade before diagnosis.

The episode begins with host Theral Timpson introducing Chris Whelan, lead author of a 2023 Nature paper from the Pharma Proteomics Project that characterized plasma proteomic profiles of 54,000 UK Biobank participants. Whelan explains that the paper was a pre-competitive effort involving 13 pharma companies, focusing on data generation and protein quantitative trait loci (pQTLs) but limited in insights due to IP concerns. He notes that over 200 downstream studies have since been published, marking a shift toward integrating proteomics in molecular epidemiology. Whelan discusses clinical translation, citing a study by Julia Carasco Zanini and Claudia Langenberg that used machine learning with protein panels to predict diseases like multiple myeloma up to 10 years in advance. He shares examples from Janssen, including using pQTLs to identify new therapeutic targets for Parkinson's disease, such as galectin-3, and a schizophrenia-related protein that is now part of a drug program. Whelan compares proteomics tools like Olink, SomaScan, and Alamar, noting their strengths and potential for clinical use. He estimates proteomics is in a phase similar to genomics in the early 2000s, with breakthroughs expected in diagnostics and disease subtyping. Whelan describes using AI on proteomic data to identify three biological subtypes of major depressive disorder based on apoptotic and inflammatory proteins. He acknowledges challenges, including the need for more diverse populations and reduced costs, but remains optimistic about the field's future.

As heard by us

A measured look at how population-scale protein data may guide future drug targets.

Large-scale proteomics comes through as a practical test of what population genetics can add to drug development. Therrell Timpson's conversation with Chris Whelan starts from the late-2023 Nature work on plasma proteomic profiles from more than 54,000 UK Biobank participants,…

Read the full review in PlayNext →

Why you'd press play

Want a clear-eyed tour of how 54,000-plus UK Biobank protein profiles can point toward new therapeutic targets?

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Mendelspod Podcast