100 million species and a trillion genes: Massive project for medical AI launched

British company Basecamp Research has announced one of the largest initiatives in biological history — the Trillion Gene Atlas project. The goal of the project is to collect and analyze genetic data from over 100 million previously unstudied living organisms. This will increase the volume of known genetic diversity on Earth by nearly 100 times. The project is being implemented with the support of Anthropic, Ultima Genomics, PacBio, and NVIDIA computing infrastructure. This is reported by Ixbt.com news.

Experts note that the combination of modern sequencing methods and AI allows for tasks that previously took over 20 years to be completed in less than two. The main task of the project is to create a completely new dataset for training biological AI models. Currently, many systems use limited open databases, and Basecamp Research believes that this data shortage is hindering the development of biological AI.

The company has already introduced its EDEN model family, trained on its BaseData database. This database contains over 10 billion genes unknown to science. According to the developers, EDEN has become the first system capable of designing therapeutic molecules directly based on disease descriptions. In laboratory tests, the model showed high activity in human T-cells and created new antimicrobial peptides against pathogens with 97% accuracy.

As part of the project, Anthropic plans to expand the scientific capabilities of its Claude AI assistant. In the future, by working with large biological datasets, Claude will help researchers interpret experimental results and develop new treatments. The Trillion Gene Atlas project is being compared in scale to the Human Genome Project, but this time it concerns a genetic map of life across the entire planet.