BioNome Assimilate Artificial Intelligence in Drug Discovery Designing and Development

Artificial intelligence is one of the colossal areas of Computer science that is mostly interested in evolving smart tools and algorithms that is competent enough to perform tasks without any human intellect. AI is progressively used in countless industries among which pharmaceutical is one of the most significant and Drug Discovery and Development is one of the leading & vital areas of research in this area. In the Drug Discovery process, the challenges are mostly time and production cost. Furthermore, the inability, imprecise delivery of the target, and insufficient dose are the hurdles that prevent the medication. Artificial Intelligence with machine learning is extremely crucial in the Drug Discovery process thus making the task easier. Technological improvements, In-silico Drug Discovery, and medication design together with advanced algorithms of AI can minimize the obstacle and impediments related to conventional drug discovery and development. Artificial Intelligence along with Machine Learning is being utilized in abundant areas of Drug Discovery that involve virtual screening, QSAR analysis, polypharmacology and many more.

More than 1060 chemical compounds are present which aids the generation of an enormous number of drugs used in pharmaceutical industries. Artificial Intelligence and machine learning can solve the problem of huge time consumption and cost in the drug discovery process that was lacking in the conventional drug discovery process. Hit and lead compounds can be determined by AI. Moreover, the confirmation and optimization of the drug target and drug structure design are way swift by applying AI. The recognition of suitable drug targets is the initial and most crucial part of the drug discovery process. Following that drug-like molecules that can interact with those targets are identified. Biomedical data repositories can also create that can be accessed in the future for any assistance.

One of the AI approaches is QSAR analysis which stands for quantitative structure-activity relationship (QSAR) analysis. It is one of the popular methods that is known for its high-throughput, high-hit rate screening. Building a QSAR model entails.

    • Gathering chemo-genomics data from a database or literature

    • Calculating appropriate descriptors from molecular representation

    • Constructing a link (model) between biological activity and the selected descriptors

    • Use of the QSAR analysis model to conclude the biological properties of several molecules

QSAR analysis rapidly predicts a huge number of physio-chemical properties like log P and log D. However, complicated features like chemical activities and side effects might not be predicted by this QSAR method. In addition to that, models based on QSAR experience issues like short training sets, inaccuracy of experimental data in training sets, lack of validation, etc. AI tools like deep learning are evolving. These tools might be used to estimate the two most crucial things which are the safety and efficacy of the compounds used for medication using big data modeling and analysis. A competition of QSAR Machine Learning was held in the year of 2012 to examine the benefits of deep learning in Drug Discovery and development process. It was found that deep learning models performed way better than machine learning approaches for a total of 15 drug candidate ADMET which is absorption, distribution, metabolism, excretion, and toxicity datasets.

QSAR analysis techniques have evolved into more advanced AI-based QSAR modeling techniques and linear discriminant analysis (LDA), support vector machines (SVMs), random forest (RF), and decision trees are examples of them. These advanced technological approaches have made QSAR analysis quick and efficient.

The analysis and illustration of the distributions of molecules along with their attributes make the virtual chemical space look like a geographical map of molecules. The objective behind this visualization of the chemical space is to gather positional information regarding the molecules inside the space for discovering bioactive molecules. Therefore, Virtual screening helps in selecting relevant molecules for testing. Examples of such chemical spaces that are open to the public are PubChem, DrugBank, etc.

Superior-profile analysis, fast removal of non-lead compounds, and therapeutic molecule selection are performed by several in silico methods that are used for the virtual screening of compounds at an affordable cost. Molecular fingerprint recognition systems and coulomb matrices are some techniques used in the drug designing process that examine characteristics like physical, chemical, and toxicological to choose the lead ingredient.

Various factors, including prediction models, molecular similarity, the process of creating molecules, and the use of in silico approaches, can be used to anticipate the desired chemical structure of a drug. A novel docking technique for receptors and 2950 ligands was introduced named DeepVS and this technique performed outstandingly when 95 000 decoys were tested against these receptors. In another work, a multiobjective automated replacement method was utilized to analyze the form similarity, biochemical activity, and physicochemical characteristics of a cyclin-dependent kinase-2 inhibitor to improve its potency profile.

            To summarise, the use of AI approaches in CADD and Drug Discovery allows for a finer knowledge about medications and health. Efficacious ways for analyzing the big data related to biomedicine aid in the identification of major targets or the definition of traits that are significantly associated with health outcomes. The current development and implementation of big data and AI approaches to build statistical and mathematical models to tackle diverse drug discovery difficulties necessitates the use of high-quality data as a key component of research.


1. Gupta, R., Srivastava, D., Sahu, M., Tiwari, S., Ambasta, R. K., & Kumar, P. (2021). Artificial intelligence to deep learning: machine intelligence approach for drug discovery. Molecular Diversity, 25(3), 1315-1360.

2. Paul, D., Sanap, G., Shenoy, S., Kalyane, D., Kalia, K., & Tekade, R. K. (2021). Artificial intelligence in drug discovery and development. Drug discovery today, 26(1), 80.

3. Lee, J. W., Maria-Solano, M. A., Vu, T. N. L., Yoon, S., & Choi, S. (2022). Big data and artificial intelligence (AI) methodologies for computer-aided drug design (CADD). Biochemical Society Transactions, 50(1), 241-252.