The COVID-19 pandemic, which was brought on by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was considered one of the worst pandemics in human history. In this article, we explore the potential of artificial intelligence (AI) in developing effective vaccines, disease treatment and flexible containment techniques.
Image Credit: creativeneko/Shutterstock.com
AI-Driven Drug Discovery
Artificial intelligence (AI), which refers to a spectrum of computing techniques, is the capability of machines to acquire new information.
The adoption of AI and its related methods, including machine learning (ML), deep learning (DL), and other traditional computational chemical tools, has substantially impacted the success level and speed of innovative pharmacological identification.
Artificial intelligence (AI) can be used to virtually assess and modify drugs, determine their bioactivities, and forecast protein-drug interactions. By generating prediction models that detect molecules with a strong likelihood of interacting with a target protein, AI systems can aid in virtual screening.
Why We Need AI for Drug Development?
The number of medications approved by the FDA is decreasing in tandem with the number of novel molecular entities (NMEs). The negative effects and decreased effectiveness of potential drugs are cited as the causes.
The pharmaceutical industry is plagued by the delay in NME production. In addition, the speculated effects of drugs on diseases are getting exponentially complicated.
AI and computational drug design offer an innovative approach to the system-centric notion for research and development leading to novel drugs, which begins with assessing the scope of a new drug.
When rapid intervention is necessary during pandemics, the design of drug-based therapy utilizing AI will be a crucial strategy to identify suitable and effective regimens.
AI and Coronavirus Drug Development
Prior to the COVID-19 pandemic, bringing drug and vaccine candidates from development to market could take between 10 and 15 years. This paradigm, however, was altered when coronavirus emerged as a serious global threat. The United States Food and Drug Administration (FDA) has only approved a few vaccines, including mRNA-1273 (Moderna) and Covishield (Oxford-AstraZeneca). However, these potential therapies still raise safety concerns; adverse events like allergic responses have been linked to their administration.
AI algorithms can be used to solve challenges with development, distribution, storage, and concerns with the safety and effectiveness of various drug/vaccine candidates.
An AI-based technology could predict drugs or peptides directly from the genomes of infected patients. As a result, they may have a stronger affinity for the target and contribute to developing novel coronavirus-targeted drugs.
This approach decreases the time and expenditure required for drug development. The molecular makeup of the drugs is specifically addressed by applications that use an AI-based approach for drug design.
Researchers worldwide have proposed novel coronavirus AI-based drugs; however, validation of these potential drugs will be required to confirm their safety and feasibility in combating coronavirus.
Recent Advances in AI-Based Coronavirus Drug Development
Recent AI applications for coronavirus include the virtual screening of repurposed drugs and novel chemical entities.
Since the human angiotensin-II converting enzyme (ACE2) is bound by the SARS-CoV-2 spike protein (S protein), this protein is a modulator of viral entrance into cells and, ultimately, infection, making it an ideal target for drug therapy.
Srinivasan et al. built a surrogate multi-task-neural network (MTNN) model to replace docking simulation in searching for new drugs targeting the spike protein.
Drugs with a high affinity for the spike protein have been employed in several reported computational investigations. These investigations, however, are only focused on the receptor ACE2/protein spike interface.
This is a restriction since it drastically decreases the probability of detecting putative allosteric antagonists of ACE2 and spike complex, resulting in the identification of a small number of potentially active drugs.
To address this, a novel ML technique SSnet was utilized to find potential drugs by screening a collection of approved drugs from the ZINC and DrugBank archives that target the ACE2 receptor in two distinct conformations (closed and open), and also ACE2 in complex with the S protein's S1 domain.
Image Credit: Mongkolchon Akesin/Shutterstock.com
The AOPEDF (arbitrary-order proximity embedded deep forest approach) program can also forecast novel drug target interactions, as demonstrated by Zeng et al.
SARS-CoV-2 3CLpro main protease (Picornain 3C-like protease), a homodimeric cysteine protease, has also been identified as a desirable target for trans-variant activity because no mutations have been identified in this protein.
An ML-based model for predicting new inhibitors was developed by Xu et al., using inhibitors of the SARS 3CLpro and COVID-19 3CLpro proteins. Their training data set contains 66 active and inactive drugs.
ISM3312 a Revolutionary AI Designed Drug for Coronavirus
Insilico Medicine stated that the China National Medical Products Administration has accepted an Investigational new drug application for a novel drug called ISM3312.
This orally accessible 3CLpro inhibitor was designed with Chemistry42 (generative chemistry platform) for the medical treatment of coronavirus. ISM3312 effectively lowered lung inflammation and viral load in tissues of the lungs in pre-clinical trials.
It also displayed broad anti-coronavirus activity, not just against SARS-CoV-2 and its well-known variations, but also against other kinds of coronaviruses, including Middle East respiratory syndrome (MERS).
Another UK-based company, Exscienti, screened 15,000 clinically validated molecules from the Scripps Research Institute in California, US, in collaboration with the Diamond Light Source, a synchrotron facility. It is also working to understand the mechanism of the coronavirus's functioning through data modeling and artificial intelligence.
Challenges and Limitations Associated with AI
AI is set to play a growing role in all aspects of healthcare. However, scaling up these designs for the real world entails numerous challenges and restrictions.
Generalization, validation, interpretability, explainability, risk-minimization, fairness, and inclusivity are a few obstacles to using AI-based drugs in healthcare and public health settings.
Future Outlook
In coronavirus research, artificial intelligence (AI) has demonstrated tremendous aptitude, from real-time virus transmission monitoring to developing new drugs and vaccines faster than before. Several molecules were recently discovered from vast databases, including billions of molecules.
The majority of those challenges are being actively addressed in the AI research community. COVID-19 has sparked a demand for going digital, enhancing data literacy, and investigating assistive algorithms in clinical settings.
In conclusion, the global pandemic has significantly accelerated the utilization of AI; therefore, it is not surprising that artificial intelligence will lead pharmaceutical development in the coming decade, speeding up the process of developing novel drugs. This will position AI at the pinnacle of the fight against public health issues. The research will not only focus on coronavirus-related issues, but also many other diseases.
References and Further Reading
Paul, D., et al. (2020). Artificial intelligence in drug discovery and development. Drug Discovery Today, [online] 26(1), doi.org/10.1016/J.DRUDIS.2020.10.010.
Exscientia | AI Drug Discovery | Pharmatech. Available at: https://www.exscientia.ai/ (accessed Sep. 25, 2023).
Anderson, A. S. (2022). A lightspeed approach to pandemic drug development. Nature Medicine, [online] 28(8), pp. 1538–1538, doi.org/10.1038/s41591-022-01945-6.
Arora, G., et al. (2021). Artificial Intelligence in Surveillance, Diagnosis, Drug Discovery and Vaccine Development against COVID-19. Pathogens, 10(8), p. 1048, doi.org/10.3390/pathogens10081048.
Kaushik, C. A. & Raj, U. (2020). AI-driven drug discovery: A boon against COVID-19? AI Open, 1, pp. 1–4, doi.org/10.1016/J.AIOPEN.2020.07.001.
Insilico Medicine: IND application for first generative AI-designed drug for COVID-19 approved. Available at: https://www.azolifesciences.com/news/20230224/Insilico-Medicine-IND-application-for-first-generative-AI-designed-drug-for-COVID-19-approved.aspx (accessed Sep. 25, 2023).
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.