Deep Learning Algorithm excavates another Biology Mystery: RNA Structure

Deep Learning falls under the category of Machine Learning, which is again a subset of Artificial Intelligence. It is also referred to as a deep neural network.

 

Deep Learning, with its incredible applications ranges from self-driving cars to entertainment and what not! Its models require little effort from the end of a programmer to concentrate on the essential features on its own.

 

(Must Check: Introduction to Deep Learning and its Applications)

 

Process

 

DeepMind has created a database of over 350,000 protein structures. By the end of 2021, the data dump is expected to have grown to 130 million buildings. It enables researchers to delve into hitherto unexplored "dark matter" of the human body's constitution.

 

As a consequence, new treatments to combat our greatest health adversaries, like cancer, are being developed, along with cutting-edge technologies to accelerate medical advances.

 

(Related Reading: Role of AI in medical field)

 

RNA didn't receive much attention until its role in the Covid-19 vaccine. Its molecules contain genetic information and can accelerate biological activities, alter your immune system, and even pass along "memories" for successive generations.

 

(Also Check: What is Genetic Engineering? Types, Process & Applications)

 

RNA, like proteins, folds into complex three-dimensional structures, but we know very little about these molecules. There are 30 times as many kinds of RNA as there are proteins, yet RNA structures have been decoded in fewer than 1% of the cases.

 

The Stanford team created ARES, a deep learning algorithm that solves RNA structures effectively. "They've made significant progress in a sector that has been resistant to transformational breakthroughs," says Dr. Kevin Weeks.

 

Despite only being trained on 18 RNA structures, the ARES robot was able to derive significant "building block" principles for RNA folding. ARES is input agnostic, meaning it isn't specialized to RNA, and may be used to resolve problems in biology, chemistry, materials science, and more.

 

According to SingularityHub, “Covid vaccine, mic drop” is perhaps the best way of putting the significance of this biomolecule in our daily life. RNA, like proteins, is transcribed from DNA and has four letters: A, U, C, and G. Killing a gene's RNA messenger is one method to turn it off without really touching it. 

 

RNA, unlike protein structures, lacks sufficient empirically tested and proven instances. ARES was trained on a limited collection of motifs from previously discovered RNA structures. It distils patterns and unique themes when confronted with a new collection of RNA structures, many of which are more complicated.

 

The structure prediction of RNA is much more difficult than proteins. For now, ARES is unable to achieve the degree of precision required for drug development or to uncover novel "hot spots" on RNA molecules. But it's a significant step in "piercing the RNA veil," according to Weeks, and one that's "poised to change RNA structure and function discovery."