Over 1,600 teams competed for a $100,000 prize in the latest Virtual Cell Challenge, according to reports from the New York Times and DNYUZ.
The Arc Institute announced on Thursday that a team from the genomics company Illumina won the first-place prize. This competition grew from last year’s inaugural challenge, which had more than 1,200 global entries.
Interest in the field is rising after massive financial commitments to biological modeling. The nonprofit research organization Biohub announced commitments of nearly $2 billion for data and modeling to help virtual cells reach their potential, the New York Times reported.
Researchers scale biological complexity
Scientists are using artificial intelligence to simulate the internal workings of human cells. While early models like GeneFormer were trained on data from about 30 million cells in 2023, current efforts are scaling much higher.
A review published in Expert Opinion on Drug Discovery said the convergence of generative AI and spatial omics is driving a move toward models that can predict cellular states and treatment responses.
Dr. Christina Theodoris, a computational biologist at the Gladstone Institutes in San Francisco, trained the GeneFormer model using data from various organs, including the lungs, eyes, and kidneys. She is training her latest model on 175 million cells. This is about six times the volume used for her previous work.
Other teams are focusing on massive data-gathering efforts like the Billion Cell Project, supported by Biohub. Dr. Alex Marson, a physician-scientist at the University of California, San Francisco, recently led a team that studied 22 million cells. His team used robotic lab technology to study these cells, systematically altering genes in 12,000 sets of immune cells to observe how they responded, the New York Times reported.
AI and biology attract investment
The intersection of AI and biology has become a high-stakes arena for startups and tech giants. Xaira Therapeutics, a company developing its own virtual cell, has secured $1 billion in investment. “A lot of companies are jumping into this space,” Dr. Xing, who co-founded the virtual-cell startup GenBio.AI in 2024, said.
But some experts urge caution regarding the speed of these advancements. Some see a revolution, while others worry about excessive enthusiasm in the field.
A.I. is going to revolutionize biology in ways we don’t yet understand.
Andrea Califano, a cancer biologist at Columbia University and head of Biohub, New York
Simon Anders expressed skepticism regarding the business side of the industry. He suggested that some may make grand promises to appeal to investors.
A few people got ahead of themselves, and other people had to rein them in.
Simon Anders
The goal of personalized medicine
Many developers want to create “digital twins” or “agentic twins” to simulate how a specific person might respond to a drug. This could allow for personalized trials in a virtual environment rather than on a patient.
Dr. Baker, a co-founder of GenBio.AI and Xaira Therapeutics, believes the next major hurdle is moving from basic models to those that can predict the behavior of entire, complex cells.
“How we build on this progress to move up the biological complexity ladder is a really important open research problem,” Dr. Baker, molecular biologist at the University of Washington, said.
Dr. Baker and his colleagues are working on engineering immune cells designed to identify and destroy cancer cells. They hope to use virtual models to test these designs before physical implementation.
The success of the team comes from a company with deep roots in the industry. Founded in 1998 and headquartered in San Diego, Illumina develops systems for analyzing genetic variation and biological function. At one point, the company’s machines accounted for more than 90% of all DNA data produced.
Researchers suggest these models are already providing value despite the long road ahead.
I think these models become useful when they become part of the playground of science.
Dr. Alex Marson