Dr. Aris Thorne, head of research at BioGenesis Pharmaceuticals, was staring at a brick wall. That wall was the stalled preclinical trial data for BGX-7, their Alzheimer’s compound. For three years, BGX-7 looked great in animal models, but human trials kept producing a mess of unpredictable responses. His in-house team was brilliant, but they’d run out of ideas and had tweaked their established protocols into the ground. The whole traditional R&D model, with its locked-down IP and siloed teams, suddenly felt less like a bastion of security and more like a cage. He knew that for BGX-7 to ever help a single patient, BioGenesis had to do something radical, they needed to embrace true open science to get their pharma innovation moving again.
Key Takeaways
- Big pharma is finally getting on board with open science. Projections show over 60% of major firms will be in at least one precompetitive collaboration by 2025.
- Using open data platforms, like the European Open Science Cloud (EOSC), can realistically slash drug development timelines by 10-15% just by giving more researchers access to the data.
- To get companies to play ball, you need clear IP rules. Things like tiered licensing agreements are key for getting firms to join consortia without giving away the farm.
- Successful open R&D is built on public-private partnerships. The Innovative Medicines Initiative (IMI) is a prime example, co-funding projects with a combined purse of over €5 billion.
- If you’re going to share data, you have to invest in cleaning it up. Strong data curation and standardization are non-negotiable if you want the data to be usable by anyone else.
Thorne’s problem at BioGenesis wasn’t unique. The entire pharmaceutical industry, for all its massive wins, was getting crushed by skyrocketing R&D costs and plummeting success rates. Back in 2024, the Tufts Center for the Study of Drug Development pegged the average cost to develop a new drug at more than $2.3 billion, and that number had been climbing for a decade. Because of that cost pressure and the sheer complexity of biology, good ideas were dying on the vine, not because the science was bad, but because the right people couldn’t see the data. “We were trying to solve a global puzzle with only half the pieces,” Thorne said later. “And we kept our few pieces locked in our own vault.”
So Thorne started looking for models outside the company’s four walls. He dug into how other scientific fields did it, especially astrophysics and genomics, where sharing massive datasets is just how work gets done. The Human Genome Project, finished ahead of schedule and under budget, was the ultimate proof of what open, collaborative research could pull off. Its principles of releasing data quickly and without restrictions had completely changed biology. But could that really work in drug discovery, an industry built on patents and cutthroat competition? A lot of people were doubtful, pointing to the huge financial risks and the need to protect every new compound.
But a few industry leaders were starting to think differently. The idea of precompetitive research was picking up steam. This meant companies would agree to work together on basic science, developing shared tools or figuring out the fundamentals of a disease, before anyone had a specific drug in hand. The goal was to de-risk the brutally expensive early stages of discovery, creating a bigger pool of validated targets and biomarkers that any company could then take and run with. The Accelerating Medicines Partnership (AMP), a public-private project started by the NIH in the U.S., was a perfect example. It brought government, industry, and non-profits together to find and validate targets for diseases like Alzheimer’s and Parkinson’s. According to an NIH press release in late 2025, AMP projects had already coughed up several new therapeutic targets for neurodegenerative diseases, saving participating companies a ton of upfront investment.
Thorne saw AMP as a blueprint. He went to the BioGenesis board with a wild proposal: instead of sitting on all their BGX-7 data, they would anonymize a huge chunk of their preclinical and early clinical trial data and contribute it to a new consortium for neurodegenerative diseases. This group, tentatively called the “NeuroDiscovery Alliance,” wanted to build a federated database where researchers from all over could analyze huge datasets of genetic profiles, imaging scans, and drug responses. The reaction was immediate. “Are you suggesting we just give away our competitive edge?” asked Eleanor Vance, the head of legal. “Our intellectual property is our lifeblood.”
This was the make-or-break moment for BioGenesis and for the adoption of open science in pharma. They had to walk a fine line between sharing and protecting their intellectual property. Thorne argued that the real value wasn’t in the raw data itself, but in the insights you could get by analyzing it all together. He laid out the challenges of one company trying to make sense of the tidal wave of biological data. “No single company, not even us, has enough computational horsepower or the right mix of expertise to squeeze every last insight out of these datasets,” Thorne explained. “By pooling our data, we get smarter, together. We aren’t giving away our drug. We’re creating a better environment for it to actually work.”
When BioGenesis finally joined, the NeuroDiscovery Alliance had a carefully built IP framework. Any data a member contributed was still attributed to them, and if a discovery came directly from one company’s proprietary data, it belonged to them. Simple. But if a new insight came from analyzing the big, aggregated pool of anonymized data from multiple sources, that was subject to a tiered licensing model. This structure meant that core insights or tools developed by the group would be available to all members on good terms, while outsiders could license them under standard commercial agreements. This hybrid model, which the Alliance detailed in a 2025 white paper, was their way of balancing scientific progress against the commercial reality of drug development, and it became the template for other groups.
For BioGenesis, one of the first big wins came from a totally unexpected place. A computational biologist at the University of Zurich, Dr. Lena Petrova, was digging through the NeuroDiscovery Alliance database. Her team used advanced machine learning to spot faint patterns in multi-omics data. When she ran her algorithms on the combined BGX-7 data alongside data from similar compounds from other companies, she found a genetic marker nobody had seen before. This marker, present in a handful of patients, was strongly correlated with bad reactions to BGX-7. It was a needle in a haystack that BioGenesis’s own team, focused on big-picture population trends, had completely missed. “It was like having a million extra pairs of eyes looking at our data, each with a different lens,” Thorne said. “We had been too close to the problem.”
With this insight, BioGenesis could completely overhaul its patient selection for BGX-7. They were now able to design a Phase II trial that specifically screened out patients carrying that genetic marker, which dramatically boosted their chances of showing the drug was both effective and safe. This saved them a lot of money, sure, but more importantly it saved them years of development time and put them on a path to get a life-changing drug to the right people. The fact that they could pivot so quickly based on an outside discovery showed the real power of an interconnected R&D ecosystem. Without that open framework, Petrova never would’ve seen the data, and BGX-7 might have been shelved for good.
The NeuroDiscovery Alliance’s success pointed to a few non-negotiables for making open science work in pharma. First, solid data governance and standardization were everything. The consortium put a lot of money into common data models and curation pipelines so that data coming in from different organizations, all collected in different ways, could actually be merged and analyzed together. The European Medicines Agency (EMA), in its 2026 guidelines on data sharing, even called for FAIR principles (Findable, Accessible, Interoperable, Reusable) to be the foundation for this kind of work. Second, you had to have trust and transparency between partners, which meant ironclad agreements on data use, publication rights, and IP. Finally, you need dedicated funding. You can’t run these massive technical platforms and coordination efforts on goodwill alone. The NeuroDiscovery Alliance pulled in major grants from government research councils and private foundations, which showed that others saw the massive public benefit in what they were doing.
The move to open science still faces plenty of challenges. Patient data privacy is a huge one that requires constant work and better anonymization tech. And let’s be honest, changing the culture inside a pharma company from a default-competitive mindset to a collaborative one takes a long time and a lot of leadership. But the benefits, like the ones BioGenesis saw with the NeuroDiscovery Alliance, are hard to argue with. You get faster discoveries, you cut down on redundant work, and you can start to tackle diseases so complex that no single company could ever solve them alone. It’s becoming a necessary path for pharma. “We still have proprietary secrets,” Thorne said at a recent conference panel. “But we’ve learned that some secrets are best shared if we actually want to cure the diseases that plague humanity.”
The future of pharma innovation is going to depend on the willingness of companies, universities, and government agencies to build these interconnected R&D ecosystems where knowledge can flow freely and strategically. This kind of collaboration promises to deliver breakthroughs faster and more efficiently than we’ve ever seen before.
So what is open science in pharma?
In pharma, open science means sharing research data, methods, and results more openly between researchers and companies. This often happens in a “precompetitive” space and can include sharing everything from preclinical and clinical trial data to biological samples and software tools, all to speed up the hunt for new drugs.
How does open science actually help R&D?
It gets more eyes on the problem. Open science fuels collaboration, cuts down on different teams re-doing the same basic work, helps identify drug targets faster, and lets scientists run more powerful analyses on bigger datasets. Having more diverse expertise looking at the data can lead to faster development cycles and better success rates for new drugs.
What are the big hurdles to making this work in pharma?
The main challenges are figuring out the intellectual property, working through data privacy laws (like GDPR and HIPAA), and getting everyone to agree on data standards so the information is actually usable. There’s also a big cultural barrier inside companies that are built on secrecy. Building trust and having clear rules of the road for everyone involved is also a must.
What’s precompetitive research and how does it fit in?
Precompetitive research is when competitors agree to collaborate on the foundational science before anyone has a specific drug candidate. It’s a key part of open science because it lets companies pool their money and people to de-risk the very expensive, early stages of discovery. This creates a stronger scientific base that they can then use for their own proprietary work.
How do you protect IP if you’re sharing everything?
You protect it with smart legal frameworks. These can include tiered licensing deals, making sure data contributions are clearly attributed, and having agreements that spell out the difference between discoveries made from one company’s private data versus insights from the combined, anonymized data. These hybrid models let you share foundational knowledge while still protecting the commercial rights to a specific drug.