AI drug discovery is only as good as the data behind it. One ALS nonprofit spent a decade building that foundation before anyone was asking for it.
At this year's Target ALS Annual Meeting, leaders from Roche, Lilly Ventures and insitro discussed applying AI to ALS drug discovery. Their tone was optimistic, with an important caveat.
AI Is Only as Good as Its Data
AI already improves some parts of drug discovery. Realizing its full potential in areas like clinical development will take more time, the companies noted.
That caveat points to a less-discussed problem. AI models only find patterns in the data they're given, so narrow or disconnected datasets limit what they can discover.
Target ALS launched its Data Engine in 2024 to address exactly that gap. It centralizes clinical, demographic and biological data, including whole-genome sequencing, proteomics and RNA sequencing, into one resource for researchers worldwide.
That's a meaningfully different approach than most disease-specific research infrastructure. Clinical and biological data usually sit in disconnected systems researchers can't easily cross-reference.
Why Representation Is Part of Data Quality
ALS affects people across every race and ethnicity, but ALS research has historically lacked that same diversity. Target ALS's Global Research Initiative is working to expand participation among historically underrepresented populations specifically.
That's not a side project separate from data quality. It's part of what data quality actually means for a disease this variable.
From Pattern to Proof
Finding a promising pattern in data is only the first step. Researchers still need to test whether it holds up biologically.
Target ALS pairs its datasets with matched biosamples, including blood, cerebrospinal fluid and postmortem tissue, available through its Research Cores. That connection lets researchers move from a computational signal to an actual experiment.
Access matters too. Target ALS provides free access to qualified researchers within 48 hours of signing a data agreement.
That removes cost and logistics barriers that typically slow this kind of research down.
"Open access and data harmonization" are major differentiators, said Jamie Ifkovits, scientific director of neurodegeneration at GSK. She said the approach helps reduce barriers between academic and industry research.
More than 660 researchers across 35 countries currently use the platform. That scale is designed to draw AI and computational biology experts into ALS research. Even those without prior experience in the disease.
The underlying bet is straightforward. Lowering access barriers brings in expertise that wouldn't otherwise focus on ALS specifically. It expands the pool of people actually working the problem.
What This Means for Miami
This is a useful model for any research institution building AI infrastructure in health care, not just ALS specifically. The same principle applies broadly: AI tools are only as valuable as the breadth and quality of data behind them.
For Miami's medical research institutions and health tech companies, Target ALS's approach offers a concrete example worth studying. Free, fast data access paired with matched biological samples accelerated research readiness here. More than any single AI model could have on its own.
AI Boom Pushes US Gas-Fired Power Building Past China