MIT researchers have incorporated a new feature into machine learning algorithms to enhance its prediction capability. ML is used as a computational tool to study the enormous data available that will help to identify potential new drugs.
The study conducted by MIT researchers is the first to put forward new molecules that can target PknB. This information is also a good starting point for drug developers to develop specific drugs. The study revealed some new leads beyond what was already known. The MIT society is determined to make a better world through education, research, and innovation. MIT’s dictum is “Mind and Hand” this signifies the combination of scholastic knowledge with practical use. This Institute is primarily committed to generate, disseminate, and preserve knowledge. Another attraction of MIT is that it fundamentally promotes and supports research that now led to implementing this technique in biology.
This is not a new technique for computer scientists; however, these are new to biology. This new technique can be applied in different fields of biology like protein design study. Many top scientists stated that this is how biology should be explored and is a standard shift. To find molecules that interact with specific targets, a few years back many tried using machine learning to study enormous databases of potential drug compounds. The drawback of this technique is that the algorithms do not perform well while analyzing molecules that are very different from those they have already seen.
The MIT researchers applied the Gaussian method to allot uncertainty values to the data that the algorithms are trained on to overcome. That way when the models are analyzing the training data, they will also regard how consistent these predictions are. In this process, the algorithm requires only a bit of training data that serves as an advantage. Another noteworthy factor of this method is that they can add any additional experimental data to the model and save it for future better predictions. Every tiny amount of data can enhance the model.
Studies also showed that this type of ML can be used to stimulate a green fluorescent protein’s fluorescent output that is used to label molecules inside living cells. Studies are also using the same to analyze mutations that drive tumor development.
