Humanity has decided to face one of the greatest scientific challenges in history, to sequence the DNA of all known life forms on the planet. The project called Earth Biogenome Project has an almost unimaginable goal, to catalog the genetic code of 1.85 million eukaryontan species in just a decade and who could make this possible, precisely artificial intelligence.

The project began in 2018 with the objective of creating a complete genetic map of the Earth's biosphere, it is a record that will serve as a basis for studies of evolution, medicine, biotechnology and environmental conservation. The task is monumental. So far only 4,386 genomes have been completely sequenced, which represents less than 1% of the goal, but advances in A and automation are accelerating the process in an unprecedented way.

The Google Research team is one of the protagonists of this revolution, in 2018 it developed Deep Variant, an AI that interprets DNA sequencing data as if they were images, an approach inspired by computer vision, instead of reading billions of DNA fragments in a linear manner, the system “sees” patterns, identifies errors and reconstructs the entire genome with almost perfect precision.

More recently, Google introduced Deep Polisher, which reduces assembly errors during sequencing by up to 50%, ensuring that important genetic variations are not lost. This is essential, since even small errors distort interpretations about diseases, mutations or evolutionary adaptations.

These tools are allowing AI to act as a digital microscopist of life, accelerating a task that previously seemed impossible. To give you an idea or reference, the human genome, which required 10 years and $3 billion to decipher in 2003, can today be sequenced in less than 24 hours for around $1,000.