AMongst the many other uses of Ai they are now using it top calculate the structure of chromatin in the nucleus of a cell. MIT chemists were able to use the data of chramitn structure that would take hours to calculate manually and used AI technology to auto map in seconds. The AI analyzes the DNA sequences with an ability to know how these sequences corespond to the chromatin stucture. This is facinating as we see aI being used in not only every day situations but also in science. This experimenting on gene's could take months though with the AI takes 20 minutes to make thousands of predictions. Not only would this alow genetiuc reseachers to further understand chromatin structure but it could help reading genomes.
A picture of a chromatin strcuture based of Dip-C study
The researchers had spent a long time using the Dip-C study to carefully break apart DNA and analyze each sequence. These DNA sequences would then lead to possible chroamtin structures which there could be hundreds of possible conformations. Previous research lead to this AI, it does these two things in order to properly calculate the chromatin structure. It reads DNA sequences which contain thousand of base pairs and then uses this sequencing in relation to known data charts to predict hudnreds of possible structures quickly. This could leaad to major develpment in the understanding of DNA as researches can study these chromatin much faster. Being able to do this so quickly could mean studying mutations in DNA, and how chroamtin structure effects cell function. With AI we are learning that information can be accessed much faster in all cases leading to unlimited benefit in this research.
https://news.mit.edu/2025/with-generative-ai-mit-chemists-quickly-calculate-3d-genomic-structures-0131
https://www.science.org/doi/10.1126/sciadv.adr8265
I recently saw a video about AI's use in predicting the structure of proteins, so it is interesting to see how AI is advancing different parts of science. Similar to its predictions about DNA, the video I watched said that a human predicting structure would take exponentially longer than AI's predictions for protein or in this case chromatin structure. Overall, I find this integration of AI in science intriguing and I am interested to see where it leads.
ReplyDelete