Can we scale our way to a better understanding of biology?
When we first had the CEO and founder of Ultima Genomics on the show, the company had just emerged from stealth with an ambitious vision. Coming from the semiconductor industry, Gilad Almogy believed that dramatically reducing the cost of DNA sequencing would enable scientists to study biology at an entirely new scale. And that scale would lead to discoveries we couldn’t yet imagine. A few years later, Ultima has announced the $60 genome and is partnering on some of the largest biological research projects ever undertaken.
Almogy returns today to discuss the company’s progress and how the arrival of AI has accelerated his original vision. Ultima’s new UG200 platform offers greater throughput, improved coverage and paired-end sequencing, expanding the technology’s applications beyond large-scale counting experiments into whole-genome sequencing and clinical testing.
Among the company’s major collaborations are the Billion Cells Project with the Chan Zuckerberg Initiative and the Arc Virtual Cell Atlas. Both aim to generate enormous datasets of cellular responses to perturbations, providing training data for AI models that could eventually predict how cells behave. Almogy argues that biology may be too complex to model using traditional mathematical approaches alone. As he puts it, “A cell is a video. It’s not an image.” He believes AI could do for biology what mathematics did for physics.
The conversation also explores Ultima’s population genomics partnerships with Mayo Clinic and Singapore. Mayo’s Precure Initiative aims to combine whole-genome sequencing and proteomics with electronic health records from a million patients, while Singapore’s national precision medicine program will test how genomic information can improve clinical care. Almogy sees these efforts as complementary to the virtual cell projects, approaching biology from the level of entire populations rather than individual cells.
Finally, Almogy discusses the growing importance of longitudinal sequencing and liquid biopsy. Ultima’s ppmSeq technology, designed to detect variants at frequencies as low as one part per million, could help enable more sensitive MRD testing in cancer treatment.
The future of biology will require models at several scales, from individual cells to entire populations, along with dynamic models that follow how human health changes over time. These models will need vast amounts of data. The hope is that AI will find insights in this data that we haven’t yet imagined, ultimately making biology more predictive and leading to earlier disease detection and better medicine.
