La Derecha Diario AlphaGenome: Google's AI that predicts mutations in human DNA
AlphaGenome, Google's new AI, can analyze human DNA and anticipate how mutations will impact it
AlphaGenome is the new artificial intelligence model from Google DeepMind that can analyze the human genome and anticipate how certain mutations affect the organism. Its ability to interpret regions of DNA that were previously a mystery makes it a breakthrough for biology and medicine.
This system was developed to understand the non-coding parts of the genome, which control when and how genes are activated. Its launch has already attracted the interest of researchers, hospitals, and universities around the world.
What is AlphaGenome and why does it matter?
AlphaGenome is an AI capable of processing up to one million genetic letters at once. Thanks to its transformer-based architecture, it can detect complex functional patterns and predict thousands of biological processes.
The most innovative aspect is that it analyzes regions known as the "dark matter" of DNA. That is where many mutations linked to diseases occur.
How it predicts mutations that could cause diseases
The AI identifies genetic functions and evaluates the impact of point mutations in DNA. For example, it managed to predict alterations in the TAL1 gene, which is key in acute lymphoblastic leukemia.
This level of analysis allows AlphaGenome to anticipate which variants could trigger diseases, without the need for prior laboratory tests.
An evolution after AlphaFold's success
AlphaGenome was created following the success of AlphaFold, another DeepMind AI that predicts how proteins fold. That development was awarded the Nobel Prize in Chemistry in 2024.
Now, AlphaGenome aims higher: it seeks to interpret the entire human genome, which has 3 billion letters. The challenge is greater because DNA functions are multiple and simultaneous.
Applications in health and diagnostics
The model makes it possible to detect the effect of a mutation on several biological processes in seconds. This can accelerate genetic diagnoses and reduce analysis time in rare diseases or uncommon tumors.
It can also be useful for predicting abnormal genetic expressions, which is key for the study of certain types of cancer.
How it works and what data it uses
AlphaGenome was trained with information from large consortia such as ENCODE, GTEx, 4D Nucleome, and FANTOM5. This allowed it to capture functional patterns in different cell types.
In practice, it can detect the start and end of genes, how RNA fragments are spliced, and which regulatory proteins are involved in each case.
More posts: