The utility of Intelligent Learning Engine in Drug Discovery Informatics

Anwar Rayan. 1Al-Qasemi Academic College, Bakka El-Garbiyye, Israel, 2RAND Biotechnologies, Kabul, Israel.

The increasing amount and complexity of molecular biology information evokes the need to focus on methods for mining this data for discovering new drugs. We aim to describe a system called “Intelligent Learning Engine Technology (ILE)”1 that mines databases represented as vectors composed of binary descriptors. ILE is an innovative prediction technique. Its Implementation enables to choose from a large number of candidate entities those entities with the largest probability to have a certain property, e.g. for a candidate molecule to be a drug for a certain disease. We review current achievements of such technology in the field of bioinformatics and chemoinformatics as well as their impacts on processes for drug discovery. The applications of ILE, such as novel lead compounds identification, virtual high throughput screening, classification of proteins and 3D structure prediction are discussed herein. Our performance study shows that ILE has increased efficiency of mining data compared to many other algorithms mostly used in the field (see ref 1).

Utilizing the ILE technology allows to generate a “molecular activity index” (MAI) based on optimized differences between bio-actives and non-active molecules. We performed virtual high throughput screening for large database of chemicals (e.g. ZINC database that contain more than 2M commercially-available compounds, http://zinc.docking.org/) and selected “focused library” with highest potential to be anticancer drug candidates. The biological activity of a new molecule (MDL1100) was assessed and compared to Paclitaxel (see table 1).

In conclusion, we show that the presented novel technology can facilitate and accelerate the process of discovering or even developing new drugs. Novel bioactive molecules against inflammation and cancer have already been discovered.

1. Anwar Rayan and Jamal Raiyn, WO/2009090613, Systems and Methods for Performing a Screening Process.