MagHelix™ AI-Based Drug Discovery (AIDD) Platform

A drastic increase in data digitalization in the pharmaceutical field has motivated the use of artificial intelligence (AI) technologies. The main application scenarios of AI to drug development include target prediction, high-throughput screening, drug design, and prediction of ADMET properties (absorption, distribution, metabolism, excretion, and toxicity). AI solutions can improve the predictability, speed, and accuracy of drug discovery, and ultimately accelerate the productivity of the entire drug development process.

 Figure 1. AI is changing drug discovery. (Illustration by Michele Marconi)

Creative Biostructure  has been adopting various strategies to actively integrate AI technologies into the drug discovery process. We have established a cross-functional team composed of biologists, chemists, data scientists, and AI experts. We can utilize your specific research data and drug target information to build models for screening and optimizing drug candidates.

Advantages of AI-based Drug Discovery

• Reduce the time it takes to discover a drug and make the research process more flexible.
• Improve the accuracy of predictions of drug effectiveness and safety.
• Improve the opportunity for diversification of drug pipelines.
• AI is a data-centric technology that has advantages in drug development in complex diseases where the target and/or mechanism are not clear, as compared with conventional computer-aided drug design (CADD)  focusing on the target and structural information.

How Our AI-based Drug Discovery Platform Works

Our team of scientists is working to integrate advanced technologies in high-quality data acquisition and machine learning algorithms related to drug discovery. As each project progresses, experimental data from each cycle allows building and improving project-specific local models.

In the process of quantitative structure-activity relationship (QSAR) analysis  based on machine learning, suitable compound data are selected to construct the training set and test set to generate reliable models. These models are applied to new compounds to construct affinity fingerprints and are utilized to train and predict models with relatively little data.

Using deep learning algorithms and computing capabilities, through the extraction, integration, and machine learning of existing compound databases, key information related to compound effectiveness and toxicity is obtained, which greatly improves the efficiency and success rate of virtual screening . Moreover, the deep neural networks (DNNs) show good performance in predicting physicochemical parameters and ADMET properties.

Creative Biostructure  provides customers with comprehensive drug discovery solutions and integrates AI technologies into drug discovery projects. By reducing the number of candidate compounds by hundreds, we can help you greatly reduce the time and cost of discovering and developing new drugs.

Reference

1. Hessler G, Baringhaus K H. Artificial intelligence in drug design. Molecules . 2018, 23(10): 2520.
2. Fleming N. How artificial intelligence is changing drug discovery. Nature . 2018, 557(7706): S55-S55.

Related Sections

• MagHelix™ CADD Platform
• QSAR Analysis
• In Silico Virtual Screening
• ADME-Tox Modeling and Prediction