AI can now predict the structure of chemical compounds

Artificial Intelligence can help the chemist in cracking the molecular structure of crystal which is much faster than the traditional modeling methods, according to the research which is published in Nature Communications yesterday.

Some of the scientist from the EPFL lab, which is a research institute based out of Switzerland, had also built a machine learning programme which is called by the name as SwiftML to predict, how the atoms in molecules shift when it exposed to a magnetic field.

Nuclear Magnetic Resonance is mainly used to work out the structure of compounds. Some of the groups of atoms oscillate at a specific frequency which also provides a tell tale sign of the number and location of electrons which each contains. But this technique is as of now not good to reveal the full chemical structure of molecules especially when the complex ones contain thousands of different atoms.

“Even for relatively simple molecules, this model is almost 10,000 times faster than existing methods, and the advantage grows tremendously when considering more complex compounds,” said Michele Ceriotti, co-author of the paper and an assistant professor at the EPFL.

“To predict the NMR signature of a crystal with nearly 1,600 atoms, our technique – ShiftML – requires about six minutes; the same feat would have taken 16 years with conventional techniques.”

Some of the researchers have trained the system on the Cambridge Structural Database, which is a dataset containing a calculated DFT chemicals shift for the thousands of compounds. As each one is made up of less than 200 atoms which include the hydrogen and carbon coupled with the nitrogen or oxygen.

SwiftML managed to calculate the chemical shifts for a molecule compound that has almost 86 atoms and the same chemical elements contain the cocaine but arranged in a different set of a crystal structure.

“This is exciting because the massive acceleration in computation times will allow us to cover much larger conformational spaces and correctly determine structures where it was just not previously possible. This puts most of the complex contemporary drug molecules within reach,” says Lyndon Emsley, co-author of the study and a chemistry professor at EPFL.

Leave a Reply

Notify of

IIoT Adoption Speeds Up with New Initiatives

Cisco Chairman Invests in Cybersecurity Startup Lucideus