The rate at which new drugs are discovered is in decline, with the number of drugs approved for use in the clinic per billion USD spent halving every 9 years since 1950. The advent of rational computer-aided drug design was heralded as a new dawn for drug discovery – with the ability to model potential drug targets, drugs and even the systems they act in, it was only a matter of time before new drugs moved swiftly through the pipeline to clinic. While these in silico approaches have certainly aided drug discovery, they have not produced the radical change in the success of candidate drugs the industry anticipated. The newest hope for revolutionizing the pharmaceutical industry comes in the form of artificial intelligence (AI)-based approaches. Big pharma has invested, and many startups are also on the AI trail to develop a technology that can make discovering new drugs faster and cheaper. With the cost of bringing a drug to market roughly estimated at around $2.6 billion and taking more than 10 years, this advance in drug discovery technology is badly needed. But is this, as computer-aided drug design was, an over-hyped time bomb waiting to be found out? Or is it truly a revolution on the cusp of realizing its promise?
What is artificial intelligence?
AI is essentially the ability of computer software to learn, rather than merely carrying out instructions it has been pre-programmed to perform. Within drug discovery this could mean looking at libraries of potential drug compounds and determining from previously successful candidates in other diseases which molecule will work for a new problem. AI comes in many forms, but machine learning and natural language processing are two which are helping drug discovery scientists find new drug candidates. Machine learning is a form of AI in which systems learn from data to determine patterns and make decisions based on these patterns, while natural language processing refers to computer systems that can ‘read’ and use written information. Professor Wlodzislaw Duch, who is Professor of Theoretical Physics and Informatics at Nicolaus Copernicus University, Poland, explains, “There has been great progress in machine learning, and it gives real hope for speeding up drug discovery. At the moment, computational modelling of drugs and their interaction sites use AI, but there is also the possibility of controlling experimental labs and running large scale experiments in an automatic way.”
How to Achieve Sub-Second Substructure Query Response Times [White Paper]
Chemical substructure and similarity searching is the cornerstone of most cheminformatics systems and a critical feature in electronic lab notebooks (ELN). Learn about a vastly more scalable and performant chemical cartridge called ChemSearch. It is used to search all chemical documents and power ChemACX® in a new cloud-based ELN.View Whitepaper
How can artificial intelligence help in drug design?
Computational power has been used to aid drug discovery since the 1980s, but before AI this mainly took the form of virtual screening, molecular modelling and predicting how likely a drug candidate is to have the right properties to be non-toxic in the body. This process essentially skims the pool of drug candidates for those most likely to be successful, meaning the later – more expensive and time consuming – steps in pre-clinical tests are not done on molecules unlikely to work.
The difference with AI is that after being trained on libraries of compounds with known properties, it can learn to make associations for itself and tell us which molecules are likely to be successful for the desired application.
David Wang, General Manager of Informatics at PerkinElmer explains, “There are three specific areas of interest where AI and machine learning can be applied; target identification, lead optimization and screening drug candidates. Target identification involves further validation and/or refinement of targetable areas for lead identification and/or compound screening. Lead optimization involves algorithms to assist in simulations, such as simulated structure, toxicity, binding and drug availability. Screening drug candidates uses algorithms for image or pattern recognition in very high content or high throughput screens to assist scientists in identifying rare or non-obvious patterns in very large data sets.”
Research organizations already using artificial intelligence in drug design
AI is no longer merely a thing of science fiction, and is already the backbone of several exciting initiatives. IBM Watson was first developed as a question answering tool and famously beat champion humans in the game show Jeopardy in 2011. IBM Watson later partnered with Pfizer in their drug discovery research in 2016. The software uses machine learning and natural language processing among other AI techniques, and Pfizer have used IBM Watson to identify new targets in their immuno-oncology research. In this case, the AI technology can help scientists sift through massive amounts of data. Along with information on the chemical properties of drug candidates and targets, IBM Watson can use information from the literature with over 25 million Medline abstracts in its system. Using this information, it can make connections a team of scientists would be unlikely to make.
GNS Healthcare partnered with Genentech, a member of the Roche group, to use machine learning for drug discovery. Their focus is on developing personalized medicine approaches using AI. A big drain in the pharmaceutical pipeline is the discovery that some drugs only work for a certain group of people – sometimes because of their genes, sometimes even their microbiome. GNS hope AI will predict which subgroups will respond to treatments to save money and enable drugs to be used only in those who will benefit from them.
Despite the promise of these AI software programs, AI-assisted treatments have not yet made to clinic. BenevolentBio is an AI startup who are hoping to change that, and have used their natural language processing software to find a drug candidate which reduced symptoms of the neurodegenerative disease amyotrophic lateral sclerosis (ALS) in mice. The platform contains information from patient records, clinical trials, the research literature and patents and can then infer relationships between potential drug candidates, diseases, genes and more to predict which drug-like molecules will work for which diseases.
Professor Duch sees AI expanding its potential in drug discovery, “One direction AI might head in is the use of quantum computers, and neuromorphic computing. Classical computing should also lead to great progress, building models of biological processes that are too complex for humans to understand in detail. For example, BioCyc (an online tool which allows analysis and prediction of genome and metabolic pathways) collects data curated from tens of thousands of publications. When you consider this, there is no alternative to AI.”
AI has not been a quick fix to solve the pharmaceutical industry’s problems. While there have been success stories, some experts wonder if it is another hype train on the road to nowhere. If scientists can continue the development of AI it stands a chance of being a paradigm-shifting technology. If not, it could merely represent another small step that helps scientists in their drug discovery quest, as computer aided drug discovery did.
Professor Duch suggests, “For AI to achieve its potential we need a combination of models based on data from literature, computational techniques and automatized labs, i.e. hardware, software and databases under the control of AI systems. Another factor is the need for better phenomics (understanding of how phenotypic traits change in relation to the environment or other factors) on many levels – the Research Domain Criteria (a computational framework integrating information from molecular to behavioral to better understand human behavior) of the NIMH are still far from being complete, so it is hard to predict how the molecular level is coupled to physiology and behavior.”
If AI technology can be developed the options are virtually limitless. A true AI technology could think like a human brain but with a much greater capacity, holding the entire system of human and disease biology in its ‘mind’ at one time, alongside all the information in the scientific literature. It could mine both these resources to find the best therapy option for any given disease and individual. Wang concludes, “Continuing advances in computer power will allow scientists to continue to build and augment existing models, further improving the adoption of synthesis planning software on a larger scale. The AI and machine learning landscape within pharma has seen massive growth within the last 5+ years, the expectation is that this growth will only continue as capabilities expand.”