Limitations of Modern Databases: Q&A with Stavros Papadopoulos

Stavros Papadopoulos
Founder and CEO
TileDB

Stavros Papadopoulos, founder and CEO of TileDB, spoke with Pharmaceutical Executive about how modern databases are struggling to handle life sciences data, which in turn limits scientific discovery.

Pharmaceutical Executive: How significant an issue is pricing and access to drugs and medical devices?
Stavros Papadopoulos: Pricing and access to drugs and medical devices is a very significant issue. A recent report found that between 2022 and 2023, prices for nearly 2,000 drugs increased faster than the rate of general inflation, with an average price hike of 15.2 percent. As a result, nearly 30 percent of Americans say they haven’t taken their medication as prescribed due to high drug prices.

Combating these price increases–and ultimately, medication non-adherence–will ultimately boil down to the ability to make drug development cycles faster and more cost-efficient. Often this means leveraging data and AI to find novel drugs, reduce failure rates and get to lower costs of innovation. In this sense, the industry needs data and AI approaches that are not fragmented and work for bleeding-edge biopharma data which is often novel, large, diverse and complex yet extremely valuable, such as genomics, transcriptomics and imaging data.

When data is a competitive advantage, a fit-for-purpose solution that supports these frontier data types and enables seamless and secure collaboration, the result is lowered software licensing costs; simplified infrastructure and a reduction in the data engineering burden from biopharma teams who can focus instead on innovating with better ROI. Ultimately, this helps drive better prices in the market.

PE: How big of an impact is digital transformation expected to have?Papadopoulos: Data is vital to digital transformation. It helps improve operational efficiency, fosters innovation, and enables life sciences organizations to make strategic and impactful decisions. Digital transformation is already a big part of the pharmaceutical industry today. Data is the key currency and life sciences organizations are looking for better ways to manage, analyze and maximize their use of it to solve big problems.

Today, data in biopharma companies often sits across multiple teams and geographies and is siloed due to software or governance and data sovereignty requirements. Bringing it all together in meaningful ways is a significant priority for digital transformation teams today. Think of drug discovery platforms for cancer that integrate genomic, clinical phenotypic, molecular and immunological data. As we have noted, traditional databases have been limited in their ability to handle this complexity and require significant engineering effort and costs.

TileDB is helping life sciences teams accelerate their journey into tapping more of their data effectively. Consider the work of Rady Children’s Institute for Genomic Medicine (RCIGM) and their BeginNGS initiative, in the area of rapid newborn genome sequencing. By utilizing huge volumes of genomics data from healthy individuals from the UK Biobank and the Mexico City Prospective Study, the BeginNGS team can now identify and remove false positive genomic variants from the BeginNGS screen. RCIGM can now effectively reduce the number of false positives in affected newborns from 97 percent to three percent–resulting in far fewer anxious families.

PE: What trends have been identified in R&D within the Pharma industry?
Papadopoulos: Biopharma teams want to take all the data they’ve accumulated, implement AI and work magic. We completely understand the “bright, shiny object” syndrome that accompanies AI projects, but these days we’re encouraging a “reverse trend”–that organizations hold off from adopting and implementing AI until they have a robust data management infrastructure in place. Without comprehensive data cataloging, structuring, collaboration, and analysis capabilities firmly in place, AI deployments will only exacerbate existing chaos.

Life sciences organizations are realizing that in many instances, they must start from scratch–building a unified, secure and discoverable data ecosystem, or a trusted research environment, before even thinking about AI. In many cases, there needs to be a complete rethinking of data infrastructure, especially to tackle technical debt of legacy systems.

PE: How are evolving customer expectations influencing strategy?
Papadopoulos: Today’s healthcare industry revolves around patients who create demand for specific drugs, including specialty drugs used to treat rare or complex diseases. This is a key element of precision medicine, and as patients and their physicians increasingly demand drugs and treatments tailored to individuals, life sciences organizations face immense pressures to bring new drugs to market quickly, safely, and cost-effectively.

AI can certainly play a key role in helping satisfy these demands, but we’re encouraged to see pharma firms taking a step back first to make sure their data infrastructure is rock-solid. They’re becoming attuned to the reality that if they don’t have a structured system to query, no amount of AI will yield meaningful insights. Only once organizations internalize this message and take active steps to get their data strategies in order, can AI move from hype to a true innovation driver.