Mapping Human Disease One Billion Cells at a Time

FEATURED RESEARCH

Why AI Drug Discovery Needs Billion-Scale Cellular Data

Over the past couple of years, AI drug discovery has been limited less by algorithms and more by biology. Models can read papers and mine databases, but they struggle when the underlying experimental data is sparse or indirect.

On January 13, Illumina announced an effort designed to change that constraint at scale: the launch of the “Billion Cell Atlas”, the first step in building a 5-billion-cell map of human disease biology.

Why scale matters in biology: Most genetic evidence used in drug discovery comes from population studies or narrow laboratory experiments.

The Billion Cell Atlas aims to show, directly and systematically, what happens inside human cells when genes are switched on or off.

Using CRISPR, Illumina will measure how 1 billion individual cells respond to genetic perturbations across more than 200 disease-relevant cell lines, covering cancer, immune disease, cardiometabolic conditions, neurological disorders, and rare diseases.

In practical terms, this means observing the effects of all ~20,000 human genes across many cell types rather than inferring biology from limited snapshots.

Built for AI training: The Atlas is being developed with founding partners AstraZeneca, Merck, and Eli Lilly and Company. Merck plans to use the data to train internal AI foundation models and build virtual cell systems that better predict disease relevance.

For AI, this dataset shifts learning away from text and toward experimentally grounded biology.

The data backbone: Illumina expects to generate around 20 petabytes of single-cell transcriptomic data per year, made possible by its single-cell RNA sequencing platform and accelerated processing pipelines.

The resulting datasets will be hosted for large-scale analysis through Illumina’s cloud infrastructure.

The bigger picture: The Billion Cell Atlas is less about one discovery and more about infrastructure.

By turning genetic signals into observable cellular behavior at massive scale, it gives AI models something they have long lacked: a dense, consistent map of how human biology actually responds to genetic change.

That foundation may shape how targets are chosen and how risk is assessed long before a drug ever reaches the clinic.

For more details: Full Article 

Brain Booster

Select the right answer! (See explanation below and source)

What Caught My Eye

DRUG PIPELINE

First Patient Dosed in Phase II Trial of AI-Developed IBD Treatment

Just this week, Insilico Medicine dosed the first patient in a Phase IIa clinical trial of Garutadustat, an AI-designed, gut-restricted drug for inflammatory bowel disease.

The study, known as BETHESDA, is evaluating the therapy in around 80 patients with ulcerative colitis across multiple centers.

Garutadustat is a prolyl hydroxylase inhibitor developed using Insilico’s generative AI platform and is designed to act locally in the gut. Rather than broadly suppressing the immune system, the drug aims to reduce inflammation while supporting repair of the intestinal barrier.

That approach targets a major limitation of current IBD therapies, where more than half of patients eventually lose response or face safety concerns.

The compound moved from AI-driven design to preclinical nomination in about 12 months, far faster than traditional timelines!

Early Phase I studies in Australia and China showed favorable safety and gut-restricted exposure.

With first patient dosing now complete, BETHESDA marks another meaningful test of whether AI-designed drugs can deliver clinical benefits in complex human diseases.

For more details: Full Article

Top Funded Startups

Byte-Sized Break

📢 Other Happenings in Healthcare AI

• AstraZeneca acquiring Modella AI brings foundation models and AI agents directly into oncology R&D, reinforcing that internal AI capability is now treated as a core competitive asset rather than an external add-on. [Link]

• European Medicines Agency and U.S. Food and Drug Administration agreed on shared principles for using AI across the full medicines lifecycle, from discovery and trials to manufacturing and safety monitoring. The move lays a common regulatory foundation for how AI-generated evidence will be evaluated on both sides of the Atlantic. [Link]

• Anthropic rolled out healthcare and life sciences tools for Claude that integrate medical records, following OpenAI’s ChatGPT Health launch. The expansion brings foundation models into patient and clinical workflows, intensifying scrutiny around safety, privacy, and oversight. [Link]

Have a Great Weekend!

Trivia Answer: D) Golgi apparatus

The Golgi apparatus processes and packages proteins and lipids that have been synthesized in the endoplasmic reticulum, preparing them for delivery to their final destinations inside or outside the cell. [Source]