Portal de Murcia

www.portaldemurcia.com

• Home
• Business
• Hotels
• Restaurants
• News
• Advertisement

The prestigious journal Applied Soft Computing has published Antonio Serrano's research, directed by professors Andrés Bueno-Crespo and José Luis Abellán, which is based on a reinforcement learning algorithm.

It is a first step in accelerating the discovery of new drugs.The prestigious Dutch magazine 'Applied Soft Computing', among the 10 best in the world in the area of ??Interdisciplinary Applications of Computer Science, has recently published a work from the doctoral thesis of researcher Antonio Serrano, of the Degree in Computer Engineering of the UCAM.

In this work, a method capable of finding new drugs up to eight times faster than traditional procedures is described.The research has been carried out by a multidisciplinary team made up of the UKEIM groups (to which Antonio Serrano, Andrés Bueno-Crespo, José Luis Abellán and Baldomero Imbernón belong), BIO-HPC (Horacio Pérez-Sánchez), and DISCA de la Polytechnic University of Valencia (José M.

Cecilia).

These researchers use a prediction method called 'QN-Docking' that uses artificial intelligence based on a reinforcement learning algorithm and a neural network that helps to find the best possible solution between potential drugs.The long times required to find and validate any drug have led in recent years to the use of Computational Chemistry, which, with techniques such as virtual screening, performs simulations and discards unviable drugs before moving on to the clinical phase.

The progress of the UCAM research is focused on the speed with which these simulations are carried out.

The 'QN-Docking' method, compared to other common procedures, offers satisfactory results up to eight times faster.

The finding, according to the researchers, is the first step in what could be a revolutionary solution to accelerate the discovery of new drugs.Through it, the algorithm executed on a computer tests the different positions of the candidate drug, describing a series of virtual trajectories until obtaining the best possible anchoring with the receptor molecule involved in a certain disease.

For now, the correct operation of this method has already been tested with the beta-cyclodextrin receptor molecule and the pharmacological candidate of kaempferol, in a simplified environment.

Source: UCAM