Google DeepMind has developed AlphaGenome, an AI model designed to interpret the enigmatic 'dark matter' of the human genome, which consists of non-coding DNA that influences gene activity. This model addresses a critical challenge in science: understanding the functions of this 98% of our genetic material, which could shed light on various diseases, including cancer. AlphaGenome's predictive capabilities mark a significant advance in genomic modeling, praised by experts for its potential to outperform existing sequence-to-function models, thereby enhancing our grasp of genetic expressions and mutations.
Nearly 25 years after sequencing the human genome, 98% of it remains a mystery; Google's AI, AlphaGenome, may unveil how non-coding DNA influences disease.
Pushmeet Kohli from DeepMind emphasizes that tackling the 'dark matter' of the genome is not only a critical problem in biology but across all scientific fields.
Computational genomicist Anshul Kundaje praises AlphaGenome as a genuine advancement over existing models, highlighting its predictive capabilities for gene expression from DNA sequences.
Unlike AlphaFold 2, which solves structural problems of proteins, AlphaGenome focuses on the complex and multifaceted roles of DNA sequences in gene regulation.
Collection
[
|
...
]