MIT researchers, led by Associate Professor Bin Zhang, have made a groundbreaking discovery in understanding the 3D structure of the human genome using artificial intelligence. The team developed a computational model called ChromoGen that utilizes generative AI to predict chromatin structures from DNA sequences. This innovation enables scientists to analyze thousands of genome structures in minutes, a significant improvement over traditional experimental methods that take weeks.
The ChromoGen model combines deep learning and generative AI to analyze DNA sequences and predict chromatin conformations. Trained on over 11 million chromatin conformations, the model can generate accurate 3D structures of genomic regions. By analyzing DNA sequences and chromatin accessibility data, ChromoGen captures sequence-structure relationships, providing valuable insights into genome organization.
This breakthrough has far-reaching implications for understanding gene expression, disease research, and therapeutic applications. By analyzing 3D genome structures, researchers can reveal how gene expression is regulated in different cell types, providing insights into cellular function and disease mechanisms. This technology could also help explore how mutations affect chromatin conformation and lead to diseases.
The researchers have made the ChromoGen model and data available to the scientific community, paving the way for further exploration of genome structure and function. With its speed, accuracy, and scalability, ChromoGen is poised to revolutionize the field of genomics and drive new discoveries in biology and medicine.
