Learning the language of molecules to predict their properties

 


Discovering new materials and drugs typically involves a manual, trial-and-error process that can take decades and cost millions of dollars. To streamline this process, scientists often use machine learning to predict molecular properties and narrow down the molecules they need to synthesize and test in the lab.

Researchers from MIT and the MIT-IBM Watson AI Lab have developed a new, unified framework that can simultaneously predict molecular properties and generate new molecules much more efficiently than these popular deep-learning approaches.

To teach a machine-learning model to predict a molecule’s biological or mechanical properties, researchers must show it millions of labeled molecular structures — a process known as training. Due to the expense of discovering molecules and the challenges of hand-labeling millions of structures, large training datasets are often hard to come by, which limits the effectiveness of machine-learning approaches.

By contrast, the system created by the MIT researchers can effectively predict molecular properties using only a small amount of data. Their system has an underlying understanding of the rules that dictate how building blocks combine to produce valid molecules. These rules capture the similarities between molecular structures, which helps the system generate new molecules and predict their properties in a data-efficient manner.

This method outperformed other machine-learning approaches on both small and large datasets, and was able to accurately predict molecular properties and generate viable molecules when given a dataset with fewer than 100 samples.

“Our goal with this project is to use some data-driven methods to speed up the discovery of new molecules, so you can train a model to do the prediction without all of these cost-heavy experiments,” says lead author Minghao Guo, a computer science and electrical engineering (EECS) graduate student.

Guo’s co-authors include MIT-IBM Watson AI Lab research staff members Veronika Thost, Payel Das, and Jie Chen; recent MIT graduates Samuel Song ’23 and Adithya Balachandran ’23; and senior author Wojciech Matusik, a professor of electrical engineering and computer science and a member of the MIT-IBM Watson AI Lab, who leads the Computational Design and Fabrication Group within the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). The research will be presented at the International Conference for Machine Learning.

Event Name : International Molecular Biologist Awards

Website Link: molecularbiologist.org

Contact Mail ID : support@molecularbiologist.org

Nomination Link  : https://molecularbiologist.org/award-nomination/?ecategory=Awards&rcategory=Awardee


Follow On:

Twitterhttps://x.com/Camilla532645                                                                   

Blogger https://molecularconference.blogspot.com/ 

Youtube https://www.youtube.com/channel/UCehrwFGWKbQa0mKDDNJCwvA                  

Pinterest https://in.pinterest.com/molecularbiologistawards/                   

Linkedin https://www.linkedin.com/feed/?trk=onboarding-landing               

Instagram https://www.instagram.com/molecularawards

facebook https://www.facebook.com/share/v/1EPhJhbUDg/

#MolecularLanguage #Chemoinformatics #MoleculePrediction #ChemicalProperties #MolecularModeling #MolecularDesign #MolecularScience #MolecularEngineering #MolecularInsights #PredictiveChemistry #MolecularAnalysis #ComputationalChemistry #MolecularInnovation #MolecularResearch #MolecularData

Comments

Post a Comment

Popular posts from this blog

Pausing” Cell Death Could Be the Key to Longevity

Image
  Necrosis drives aging and disease by triggering damaging inflammation. Interrupting it may offer new treatments for chronic conditions and improve health in space travel. Necrosis, a form of uncontrolled cell death, may hold one of the most promising keys to altering the course of human aging, disease, and even space travel, according to a new study by researchers at UCL, the drug discovery company LinkGevity, and the  European Space Agency  (ESA). Challenging conventional thinking, the paper draws on evidence from cancer biology, regenerative medicine, kidney disease, and space health to argue that necrosis is not just a final stage of cell death, but a major driver of aging—and a potential target for intervention. Dr. Keith Siew, an author of the study from UCL Centre for Kidney & Bladder Health, said: “Nobody really likes talking about death, even cell death, which is perhaps why the physiology of death is so poorly understood. And in a way, necrosis is death. If...
Read more

Record-Shattering Molecule Stores Data at “Dark Side of the Moon” Temperatures

Image
  Researchers from The University of Manchester and The  Australian National University  (ANU) have developed a novel molecule capable of storing data at extremely low temperatures, comparable to the frigid conditions on the moon’s dark side at night. Their results, published in  Nature , point toward the potential for future data storage devices no larger than a postage stamp, yet capable of holding up to 100 times more information than today’s leading technologies. “The new single-molecule magnet developed by the research team can retain its magnetic memory up to 100 Kelvin, which is about minus 173 degrees  Celsius , or as cold as an evening on the moon,” co-lead author Professor Nicholas Chilton, from the ANU Research School of Chemistry, said. “This is a significant advancement from the previous record of 80 Kelvin, which is around minus 193 degrees Celsius. If perfected, these molecules could pack large amounts of information into tiny spaces. “Pink Floyd’...
Read more

Does cellular senescence hold secrets for healthier aging

Image
  Disease, injury, and other stress factors harm cells throughout our bodies. Ideally, the damaged cells are cleared by our immune systems through a process called apoptosis. But as we age, our bodies are no longer as effective at removing dysfunctional cells, and this can contribute to a weakened immune system and other less efficient biological processes. An increasing number of researchers are exploring whether learning to harness a cellular state known as senescence — during which damaged cells resist removal by apoptosis, linger, and harm neighboring normal cells — might hold the key to revitalizing aging tissues and increasing healthy, active years of life. Senescent cells are unique in that they eventually stop multiplying but don’t die off when they should. They instead remain and continue to release chemicals that can trigger inflammation. Like the one moldy piece of fruit that corrupts the entire bowl, a relatively small number of senescent cells can persist and spread in...
Read more