Artificial Intelligence — The future of pharma industry

AI is going to play a critical role in pharmaceutical industry to explore the unmet medical needs of healthcare sector, to meet the pace with which resistance is being developed for molecules and to match the rate at which new diseases are being identified, informs Pirthi Pal Singh, Associate Director, Dr Reddy’s Laboratories

The term artificial intelligence (AI) was coined somewhere in 1950s, but has become a buzzword in the last few decades, thanks to advanced algorithms and increased volumes of raw data. AI applications have increased exponentially and have become an integral part of our life [1].

Primarily, AI is an information-processing paradigm inspired by the way a mammalian brain processes information. The power of AI is that it learns from the historical data. A variety of mathematical models illustrating the functional aspects of the brain’s basic element, the neuron, have been developed. Such mathematical models are termed as artificial neural networks (ANNs), parallel distributed processing (PDP), adaptive systems, connectionist networks, etc.

Role of AI in pharma industry

AI is going to play a critical role in pharma industry to explore the unmet medical needs of healthcare sector, to meet the pace with which resistance is being developed for molecules such as anti-tuberculosis, and to match the rate at which new diseases are being identified. The traditional drug discovery approach can take upto 10 to 15 years and about $2.5 billion investment to bring a molecule from conceptual stage to market [2,3]. In summary, lower success rate in drug discovery phase with huge investment of money and time, is one of the most critical reasons for decline in number of NCEs being discovered. In order to combat these challenges, many pharma companies have already adopted AI in their research programme. Following sections below illustrate how different pharmaceutical companies are engaging AI-based platforms at various stages of their research programme [4,5,6].

AI in drug discovery

Many pharma companies, in collaboration with AI companies, have developed cloud-based AI platforms to accelerate their drug discovery programmes. These platforms look for pattern in data and make use of algorithms that can make accurate predictions about the potential drug molecules based on computational structure analysis, drug target and data from in-vivo cell line studies [1]. For instance, Watson Health and Pfizer announced a collaboration to accelerate Pfizer’s immuno-oncology discovery programme using cloud-based Watson’s machine-learning system. The IBM Watson platform will aid in identification of new drug targets, fixed-dose combinations to be studied and provide assistance in selecting patients for trials [7].

Similarly, UK-based AI drug discovery company Exscientia signed one of the biggest AI drug discovery deal with Celgene to accelerate its small molecule discovery programme in oncology and autoimmunity segment [8]. In addition to it, Exscientia is also supporting drug discovery programme of Sanofi, GlaxoSmithKline, Sumitomo, Evotec, etc, using its artificial intelligence algorithms [9,10].

AI in genomics

Verge Genomics has developed a platform technology that maps genes which are responsible for causing disease and then maps the drug molecule that target them to provide cure. Currently, the platform is being used to discover molecules for treatment of neurological diseases [4].

AI in diagnosis

FDA granted Bayer and Merck & Co with Breakthrough Device Designation for AI pattern recognition software that analyses images from cardiac, lung perfusion and pulmonary vessels [11]. This software will support radiologists by identifying signs of Chronic Thromboembolic Pulmonary Hypertension (CTEPH), a rare form of
pulmonary hypertension.

AI in drug repurposing

Drug repurposing is an application of approved drugs for the treatment of a different disease. AI platforms are a boon for drug repurposing, wherein the available data of drug molecules is evaluated to match new targets. HealNet is one of the largest and complex database systems available for existing drugs for rare diseases. This database is developed by Healx and it contains more than billion documented interactions among patients, existing drug molecules and rare diseases. It uses machine learning and AI to repurpose drug molecules for curing rare diseases. Also, Ligand Express, a cloud-based platform from Cyclica, leverages biophysics and AI to identify drug target, mechanism of action, elucidation of adverse effect and repurposing of small molecules [4].

AI in personalised medicine

All individuals are not same with respect to physical structure, rate of metabolism, genetic makeup etc, and therefore the therapy/dose needs to be personalised based on individual requirement. Methods like artificial intelligence and the underlying machine learning can provide the framework to stratify patients, initiate specific tailored treatments and thus increase response rates, reduce adverse effects and medical errors. GNS Healthcare’s “Reverse Engineering and Forward Simulation” (REFS), a machine-learning and simulation platform, aids in finding and validating potential new drug candidates based on patient response marker and thus leading to personalised treatments that are better match to individual patients [12].

AI in drug product development

Self-learning AI platforms like Artificial Neural Network (ANN) and Design of Experiment (DoE) helps in understanding inter-parameter interactions and further supports in composition and process optimisation. It helps in developing a multivariate correlation to obtain a quality drug product, based on understanding of cause-effect relationship between formulation ingredients/process parameters and quality target product profile [13].

AI in clinical monitoring

Tencent Holdings, along with Medopad, has developed AI algorithms for patients suffering from Parkinson’s disease. AI monitors patient’s movement via smartphone camera and determines the severity of the symptoms. Further, it also permits the doctor to monitor patient remotely, adjust dose and fix doctor’s appointment.

AI in medical development

Both Alzheimer’s and Parkinson’s patients are reported to accumulate toxic proteins that results in impaired brain functioning and leads to death of nerve cells. Mission Therapeutics, along with AbbVie, has developed an enzyme platform that enables the degradation of toxic proteins and prevents their accumulation by modulating specific deubiquitylating enzymes [14].

Current challenges of AI application in pharmaceutical industry

AI in the near future is going to play a vital role in pharma  industries. It is an effective tool for reducing investment and time to bring molecules from discovery phase to development and finally to market. Its potential to increase the success rate in drug discovery phase has already been established. However, a majority of the pharma industry still processes the data in a conventional way, although AI and machine-learning platforms are available since decades. There is a need to upgrade IT infrastructure and also a shift is required in the mindset of researchers to believe inthe power of AI. In this regard, an industry-academia collaboration can help to run pilot projects and thereby enhance acceptability of AI.

References
1. Artificial Intelligence What it is and why it matters.
https://www.sas.com/en_us/insights/analytics/what-is-artificial-intelligence.html
2. Computer-calculated compounds. Nic Fleming. Nature, Vol 557, 31 May 2018, S55-S57.
https://www.nature.com/magazine-assets/d41586-018-05267-x/d41586-018-05267-x.pdf
3. How artificial intelligence is the future of pharma. Jackie Hunter, BenevolentBio. Drug Target review, 5 December, 2016.

How artificial intelligence is the future of pharma


4. Artificial Intelligence & the Pharma Industry: What’s Next. Codrin Arsene.

Artificial Intelligence & Pharma: What’s Next?


5. AI in Pharmaceuticals. Donna Conroy and Michael Conroy.
http://www.pharmexec.com/ai-pharmaceuticals
6. 129 Startups Using Artificial Intelligence in Drug Discovery. Simon Smith.
https://blog.benchsci.com/startups-using-artificial-intelligence-in-drug-discovery#step8
7. IBM and Pfizer to Accelerate Immuno-oncology Research with Watson for Drug Discovery.
https://www.pfizer.com/news/press-release/press-release-detail/ibm_and_pfizer_to_accelerate_immuno_oncology_research_with_watson_for_drug_discovery
8. AI drug R&D company Exscientia signs deal with Celgene.

AI drug R&D company Exscientia signs deal with Celgene


9. GlaxoSmithKline signs $43m deal with AI start-up. Rachel Connolly.

GlaxoSmithKline signs $43m deal with AI start-up


10. AI-Based Drug Discovery Biotech is Recruited by Sanofi in €250M Deal.

AI-Based Drug Discovery Biotech is Recruited by Sanofi in €250M Deal


11. FDA Grants Breakthrough Device Designation for CTEPH Pattern Recognition Artificial Intelligence Software from Bayer and Merck.
https://www.prnewswire.com/news-releases/fda-grants-breakthrough-device-designation-for-cteph-pattern-recognition-artificial-intelligence-software-from-bayer-and-merck-300758747.html
12. GNS Healthcare Announces Collaboration to Power Cancer Drug Development with REFS™ Causal Machine Learning and Simulation AI Platform.

GNS Healthcare Announces Collaboration to Power Cancer Drug Development with REFS™ Causal Machine Learning and Simulation AI Platform


13. Quality by Design Approach: Application of Artificial Intelligence Techniques of Tablets Manufactured by Direct Compression. Buket Aksu,corresponding author Anant Paradkar, Marcel de Matas, Özgen Özer, Tamer Güneri, and Peter York. AAPS PharmSciTech. 2012 Dec; 13(4): 1138–1146.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3513460/
14. Mission Therapeutics and AbbVie sign DUBs Collaboration in Alzheimer’s and Parkinson’s Disease.

Mission Therapeutics and AbbVie sign DUBs Collaboration in Alzheimer’s and Parkinson’s Disease