Scientists discovering new drugs are looming large: the estimated US$2.6 billion price tag of developing a treatment. A lot of this effectively goes down the drain, as it involves money spent on nine out of ten candidates that fail somewhere between Phase I trials and regulatory approval. Some in the field are skeptical of the need to do things differently.
Leading biopharmaceutical companies believe a solution is at hand. Pfizer is using IBM Watson, a system that uses machine learning, to power immuno-oncology drug discovery.
Sanofi has signed an agreement to use UK start-up Exscientia’s artificial-intelligence (AI) platform to discover metabolic-disease treatments, and Roche subsidiary Genentech to help run the multinational company in Cambridge, Massachusetts. I am using AI system from GNS Healthcare. Search for cancer treatment. Most of the big biopharma players have similar collaborations or internal programs.
If the proponents of these technologies are right, AI and machine learning will usher in an era of faster, cheaper and more effective drug discovery. Some are skeptical, but most experts expect these tools to become increasingly important.
This change presents both challenges and opportunities for scientists, especially when technologies are combined with automation (see ‘Here come robots’). Early career researchers, in particular, need to catch up on what AI can do and how best to acquire the skills needed to be employable in tomorrow’s job market.
AI pioneers of the 1950s discussed building machines that could sense, reason and think like humans – a concept known as ‘general AI’, which has remained in the realm of science fiction for some time. is likely to.
However, over the past two decades the continued rapid increase in computer-processing power, the availability of large data sets, and the development of advanced algorithms have led to major improvements in machine learning. This has helped bring about ‘narrow AI’, which focuses on specific tasks.
These include improved abilities to analyze, understand, and generate text and speech through AI technology called natural-language processing, and artificial neural networks designed to mimic the way our brains make sense of the world. .
Such techniques are already in widespread use in areas such as computer vision, voice analysis and route selection. This advancement has also triggered a wave of start-ups that employ AI for drug discovery, many of which use it to identify hidden patterns in vast amounts of data.
For example, researchers at biotechnology company Berg near Boston, Massachusetts, have developed a model to identify previously unknown cancer mechanisms using tests on more than 1,000 cancer and healthy human cell samples. They modeled diseased human cells by varying sugar and oxygen levels, and then tracked their lipid, metabolite, enzyme and protein profiles. The group uses its AI platform to generate and analyze massive amounts of biological and outcome data from patients to uncover key differences between diseased and healthy cells.
Berg’s approach aims to identify potential treatments based on the exact biological causes of the disease. Niven Narayan, co-founder and chief executive of Berg, says, “We are reversing the drug-discovery paradigm to derive more predictable hypotheses using patient-driven biology and data rather than the traditional trial-and-error approach. ”
Using this approach, Narayan’s team identified the importance of certain naturally occurring molecules in cancer metabolism. This prompted the group to explore how a new cancer drug worked, and indicated some potential therapeutic uses. The drug, BPM31510, is currently in a phase II clinical trial involving people with advanced pancreatic cancer. The company is also using this AI system to find drug targets and treatments for other conditions, including diabetes and Parkinson’s disease.
London-based start-up firm BenevolentBio has its own AI platform, into which it feeds data from sources such as research papers, patents, clinical trials and patient records. It creates a representation based in a cloud of over a billion known and predicted relationships between biological entities such as genes, traits, diseases, proteins, tissues, species and candidate drugs.
This can be queried like a search engine, for example, to produce a medical condition and the genes associated with it, or a ‘knowledge graph’ of compounds shown to affect it.
Much of the data that the platform crunches is not annotated, so it uses natural-language processing to recognize entities and understand their links to other things. “AI can put all of this data into context and offer the most important information for drug-discovery scientists,” says Jackie Hunter, chief executive officer of BenevolentBio.