Artificial Intelligence: The cure to ‘unconquerable’ health challenges

Antibiotics are medications that destroy or decelerate the growth of bacteria and have been consumed worldwide to fight bacterial infections. Their misuse has led to increasing antibiotic resistance, where bacteria grow capable of resisting the effects of antibiotics. Based on this growing demand for new antibiotics to combat the race between mankind and antibiotic resistance, MIT researchers have successfully employed the advancing technology of Deep Learning, a branch of Artificial Intelligence, to create a model capable of identifying novel antibiotics by screening large collections of chemical compounds in a matter of days. This has led to the discovery of Halicin, a powerful antibiotic that kills several strains of deadly bacteria which have already developed resistance to existing antibiotics. This intertwinement of modern technology and medicine leads mankind towards many futuristic, revolutionary approaches that are capable of conquering some of the most concerning health challenges in the 21st century. 

The Demand for novel Antibiotics 

The need to develop new antibiotics to fight the ever-growing antibiotic resistance worldwide remains as one of the imperative health problems that modern medical scientists must strive to solve; it was declared as one of the top 10 global public health threats facing humanity by WHO. But why exactly does it pose such a threat to global health? 

Development of antibiotic resistance; Credit by DES Daughter  https://flic.kr/p/rU8JWJ

Ever since penicillin was discovered as the first antibiotic in human history by Alexander Fleming in 1928, antibiotics have revolutionised the practice of medicine. They have saved millions of lives, but the rapid spread and the increased frequency of their consumption have escalated the birth of drug-resistant strains of bacteria, ‘superbugs’, and endangered the drugs’ efficacy. It was estimated that without the invention of new antibiotics, antibiotics’ resistance could result in an annual death of 10 million people worldwide by 2050. More concerningly, bacteria’s acquisition of drug resistance seems to be outpacing the human ability to develop new antibiotics, as many researchers have predicted antibiotic resistance as a ‘losing battle’ between humans and disease-causing bacteria. In a research paper published in 2013 in the New England Journal of Medicine, the author, Spellberg from UCLA, has described the situation as “living in a bacterial world where we will never be able to stay ahead of the mutation curve”. Unfortunately, this description seems to have come true: no new classes of antibiotics have been discovered since the 1980s and it could take around 10 to 15 years and the investment of over one billion dollars to develop a new kind; the science behind antibiotics’ development is difficult and involves a significant input of time and money. 

This, therefore, links us back to our core discussion of modern technology and medicine. Will the incorporation of modern technology into the development of medical solutions be the next cornerstone that advances us into a new age of better health? 

The Deep Learning Algorithm 

Deep Learning is a part of the broad family of Machine Learning and is defined as an AI function that simulates the architecture of the human brain; it is based on the ability to learn, make decisions and improve independently without human supervision. What differentiates it from original Machine Learning is its association with artificial neural networks: computer systems that are designed to mimic the way that the human brain analyses and processes information. 

Using a model based on Deep Learning algorithm, MIT researchers have identified a powerful antibiotic that has proven to be fatal against many existing superbugs such as Acinetobacter baumannii and Enterobacteriaceae, two of the three high-priority pathogens that have been ranked as “critical” for new antibiotics to target by WHO. The antibiotic was named Halicin after the murderous AI antagonist Hal from the film 2001: A Space Odyssey, which symbolises the drug’s strong killing ability against bacteria. What is exceptional about Halicin is its ability to target against bacteria using different mechanisms in comparison to traditional, existing antibiotics, making it much more difficult for bacteria to develop resistance to. The novel drug was tested against 36 different panels of drug-resistant pathogens including the ones that are multidrug resistant and was found to be effective against 35 of them. This finding was referred to as “one of the more powerful antibiotics that has been discovered to date” by James Collins, a bioengineer on the team of MIT. Collins also described his team’s mission as wanting “to develop a platform that would allow us to harness the power of artificial intelligence to usher in a new age of antibiotic drug discovery”. 

So how exactly did the Deep Learning algorithm discover Halicin? The model was first trained with a diverse set of 2500 chemical compounds to identify molecules that are effective against E.Coli, a dangerous pathogen that has been developing resistance to many critically important antibiotics since 1950. After training, the model was then tested on a library of more than 6000 compounds with the purpose of discovering drugs that work in complete different mechanisms to existing ones. This has led to the discovery of Halicin, the one molecule that was predicted to have strong antibacterial activity and a chemical structure that is unlike any existing antibiotics. It merely took the algorithm a matter of hours to assess the compounds and identify potential antibiotics. Jonathon Stokes, the first author of the study, suggested that this “dramatically reduces the time and cost to look at the compounds” because these experiments were performed on computers. The team then progressed to a larger scale where they set the algorithm to work on a library of 107 million compounds, and a shortlist of 23 potential antibiotics was later returned after only 3 days. This discovery was again described as “remarkable” by researchers from other institutions, as it would be impossible to physically test over 100 million compounds for antibiotic discovery using traditional antibiotic screening methods, which are usually significantly costlier, more time-consuming and limited to a narrow spectrum of chemical diversity. 

This successful integration of AI technology and medicine reinvigorates the future of antibiotic drug discovery as it offers mankind a cheap, efficient alternative to traditional methods and rescues the already lagging industry of antibiotic development, which has been suffering due to its unprofitable nature in comparison to other drugs. Furthermore, it was also revealed that the same approach could be applied to the development of other drugs, such as those associated with the treatment of cancer or neurodegenerative diseases. 

Wider implications 

More generally, this combination of technology and biomedical sciences is commonly referred to as Bioengineering, a relatively new, interdisciplinary field that combines the principles of biology, medicine and engineering to create tangible, cost-efficient solutions which aim to resolve some of the greatest challenges in the modern healthcare industry. Some of the branches include: 

  • Molecular and cellular engineering such as CAR-T cell therapy, where immune cells (namely T cells) from the human body are genetically engineered to express specific receptors so that they are able to recognise cancer cells and effectively target and destroy them. 
  • Tissue engineering and regenerative medicine, which involves the practice of combining biomaterial scaffolds, cells and biologically active molecules into functional tissues in order to restore, maintain or improve damaged tissues or whole organs. Current examples include artificial skin and cartilage. 
  • Neurotechnology, which involves the development of techniques to exploit the properties of human neural systems in order to create brain-computer interfaces. Such inventions are associated with many useful applications, such as neurorehabilitation and the application of Artificial Intelligence in medical solutions
Brain-computer interface involved in the use of mind-controlled neuroprosthetics; Credit by Ars Electronica https://flic.kr/p/bGp4te

In conclusion, it is undeniable that many of the most recent, groundbreaking medical discoveries and advancements are associated with the application of modern technology and engineering principles. This pioneering, interdisciplinary fusion of the two long-existing fields extends human capabilities beyond physical, biological limits and guides us towards a blooming future of longer lives and better health. 

Leave a Reply