A Structure-Based Drug Discovery Paradigm

Abstract

Structure-based drug design is becoming an essential tool for faster and more cost-efficient lead discovery relative to the traditional method. Genomic, proteomic, and structural studies have provided hundreds of new targets and opportunities for future drug discovery. This situation poses a major problem: the necessity to handle the "big data" generated by combinatorial chemistry. Artificial intelligence (AI) and deep learning play a pivotal role in the analysis and systemization of larger data sets by statistical machine learning methods. Advanced AI-based sophisticated machine learning tools have a significant impact on the drug discovery process including medicinal chemistry. In this review, we focus on the currently available methods and algorithms for structure-based drug design including virtual screening and de novo drug design, with a special emphasis on AI- and deep-learning-based methods used for drug discovery.

Keywords: artificial intelligence; deep learning; neural network; scoring function; structure-based drug discovery; virtual screening.

Figure 1

A workflow diagram of structure-based drug design (SBDD) process. The first panel shows the human genome sequencing followed by extraction and purification of the target proteins. Second panel represents the structure determination of the therapeutically important proteins using integrative structural biology approaches. Third panel represents the database preparation of the active compounds. The next step is identification of the druggable target protein and its binding site. Subsequently, the databases of active compounds are screened and docked into the binding cavity of the target protein. In the last panel, the identification of the potent lead compound is shown. The top hit compounds obtained as a result of virtual screening and docking are synthesized and tested in vitro. Further modifications can be done for optimization of the lead compound.

Figure 2

The interaction diagram of drugs identified by SBDD methods, with their respective therapeutic targets. (a) An interaction of raltitrexed with thymidylate synthase (Protein Data Bank (PDB) ID: 5X5Q). (b) An interaction of amprenavir with HIV protease (PDB ID: 3EKV). (c) Isoniazid, a drug for tuberculosis, identified by the SBVS method (PDB ID: 1ENY). (d) Pim-1 kinase inhibitor, benzofuropyrimidine, for the treatment of various types of cancers (PDB ID: 4ALU). (e) Epalrestat is an aldose reductase inhibitor (PDB ID: 4JIR). (f) Flurbiprofen is a cyclooxygenase 2 inhibitor (PDB ID: 3PGH).

Figure 3

A workflow of the generative adversarial network approach with an artificial neural networks (ANN) for new molecule design.