AI is sequencing every species on Earth – here’s how

According to New Atlas, the Earth BioGenome Project is an ambitious $5 billion effort to sequence the genomes of all 1.85 million known eukaryotic species on Earth by 2028. The project began in 2018 and involves over 60 global affiliated projects, but has only completed 4,386 species genomes so far. Google Research has developed AI tools like DeepVariant and DeepPolisher to accelerate the sequencing process, with DeepPolisher reducing assembly errors by 50%. The technology is already showing real-world impact, helping conservationists increase the endangered Kākāpō parrot population from just 49 birds in 2015 to over 250 today. Dozens of species go extinct daily, making this genomic cataloging crucial for understanding and preserving biodiversity.

Why this matters now

Here’s the thing – we’re living through what scientists call the sixth mass extinction. Species are disappearing at an alarming rate, and once they’re gone, we lose their genetic blueprint forever. The Earth BioGenome Project is basically creating the ultimate biological library at the exact moment we need it most. And the timing couldn’t be more critical – the project started in 2018, but we’re already years behind on biodiversity targets globally.

Google’s AI game changer

What’s fascinating is how AI is transforming what used to be an incredibly slow, expensive process. Remember that first human genome? Took a decade and cost $3 billion. Now we can sequence a genome in a day for about $1,000. Google’s DeepVariant turns genome assembly into an image recognition problem, while DeepPolisher cuts errors in half. These aren’t just incremental improvements – they’re the kind of breakthroughs that make previously impossible projects actually feasible.

Real conservation impact

The Kākāpō story shows this isn’t just academic. Conservationists used genomic data to identify where genetic diversity was strongest among the remaining 49 birds, then strategically relocated them to encourage breeding. The result? The population has quintupled in less than a decade. That’s the power of having the actual genetic blueprint – you can make smarter decisions about breeding programs, habitat management, and conservation priorities. It’s like having the instruction manual for saving a species.

The scale challenge

Now, let’s be real – 4,386 genomes down, 1.85 million to go? That’s less than 0.25% of the target. The project has a massive scaling problem. But the trajectory of sequencing technology gives me hope. Costs are plummeting while speed and accuracy are skyrocketing. The teams behind projects like EBP-affiliated efforts are essentially building the infrastructure for the future of biology. What happens when we can sequence species in the field in real-time? Or use AI to predict how climate change will affect genetic diversity? We’re at the very beginning of a revolution in how we understand and protect life on Earth.