Guest Column | May 7, 2025

PheWAS Is Your Friend

By Stephen Pinkosky, vice president of drug discovery and early preclinical development, Esperion

abstract heart of pixels-GettyImages-917715400

Remember those promising new acquaintances you made a decade ago? They had unusual names, such as “PheWAS” or other mouthfuls that ended in “-omics.” Others in the industry spoke well of them for a while, but you never managed to get together and really click. Well, they're back, and those of us involved in drug development and patient advocacy are taking them seriously.

The marriage of genome sequencing and Big Data promised a great deal when it first emerged in the 2010s.  Scientists expected rapid improvements in disease prediction through revelations about the molecular “initiating factors” that set diseases in motion. Drug developers hoped for quick identification of new therapeutic targets and biomarkers, reduced attrition of drug candidates, and a better ability to track treatment outcomes.

In practice, however, it proved to be a classic example of Microsoft founder Bill Gates’s observation about technology development writ large: “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.”

Phenome-wide association studies (PheWAS) use large data sets to search for phenotypes — observed traits such as a disease state — associated with a specific genetic variant. PheWAS, as well as similar data-analysis and screening tools, are not new but have recently begun to bear some of their promised fruit. They needed some test runs and more data than existed in years past, from laboratory experiments, clinical trials, and those ever-so-slow-to-emerge patient medical records.

With data now flowing and experience piling up, the transformational impact of these tools may soon make up for lost time. Already, insights from “multi-omics” — the networked relationships of genome, proteome, transcriptome, and so forth — and the use of machine learning (ML) enable researchers to stratify risk in patients as never before and to determine their progress across a disease spectrum. This is good news for drug hunters trying to identify a manageable number of viable targets from the vast array of genetic clues. It is also good news for patients, especially those who must contend with rare diseases or don’t respond to existing treatments. 

At Esperion, my colleagues, scientific partners, and I see the possibilities firsthand, and a brief summary may inspire others. In our case, the UK BioBank provided the trove of data we needed on liver disease in 500,000 adults, with rich detail on phenotypic and genotypic variations for a wide range of health outcomes. Having experience with the ATP-citrate lyase (ACLY) protein’s role in metabolism, we utilized PheWAS to identify associations between the genetically predicted inhibition of ACLY in certain patients and the potential protective effects of such inhibition against liver diseases.

The results encouraged us. PheWAS revealed consistent associations of genetically predicted ACLY inhibition with significant reductions in the risk of neoplasms such as liver cancer, aortic atherosclerosis, and liver disorders. PheWAS also suggested that ACLY inhibition could decrease the risk of so-called “composite hepatic events” (CHEs) associated with impaired liver function.

Multi-omics analyses took us even further, helping us validate existing disease pathways and identify new ones within the immense complexity of the metabolic system.

The result is growing confidence that a drug candidate in our pipeline could improve primary sclerosing cholangitis (PSC), a disease characterized by inflammation and fibrosis of the liver and bile ducts, ultimately leading to liver failure. PSC is considered a rare disease, though it has more than 40,000 American sufferers. They have only 10 to 20 years between diagnosis and the need for a liver transplant to prevent death.

We will have a lot of work to do on our candidate, and drug development comes with no guarantees. However, this story offers three key takeaways for anyone concerned about progress in human health.

First, PheWAS and multi-omics are massive accelerators compared to earlier discovery mechanisms, which largely depended on inference from preclinical models or simply on intuitions arising from biology and pathophysiology. It is like the difference between looking for a needle in a haystack with your bare hands and using a powerful metal detector.

Second, PheWAS and multi-omics help to level the development playing field a bit. Esperion is not a cash-rich Big Pharma, but we had the resources to find and follow the data to some critical insights. Harnessed to such breakthroughs, the future of small biotechnology firms may be brighter than some observers think. 

Third, peak PheWAS and multi-omics have not been reached. The datasets continue to grow in size and quality. Collaboration is improving among a critical mass of experts in academia, medicine, and industry to make everyone’s data scans more effective. And the rise of ML and artificial intelligence is a godsend. The more “multi” the “-omics,” the more important the computing power.

We’ve crossed the underestimation threshold on Mr. Gates’ 10-year continuum. Let those old data-crunching acquaintances back in, and get ready for some serious transformation.  

About The Author:

Stephen Pinkosky, Ph.D., is the vice president for drug discovery and development at Esperion Therapeutics, which he joined in 2008. Prior to Esperion, he served as a research and development scientist at Aastrom Biosciences and as a specialist in vascular biology and inflammation in two divisions of Pfizer. Pinkosky specializes in the strategic leadership of innovative drug discovery programs that result in the filing of INDs. He holds a Ph.D. in nutrition and metabolism from the School of Medicine at McMaster University in Canada and a Master of Science degree in molecular, cellular, and developmental biology from the University of Michigan.