Finding Innovative Treatments for COVID-19 Through AI Research
Written on
Chapter 1: Introduction to AI in Medicine
Artificial Intelligence (AI) is making significant strides in the medical field, especially in the quest for effective treatments for COVID-19. Insilico Medicine, a U.S.-based startup, is leveraging machine learning technologies to identify potential molecules that could lead to viable treatments.
During a recent gathering of the World Health Organization (WHO), the outbreak initially termed "coronavirus" was officially renamed "COVID-19." This change was made to prevent the stigmatization of specific regions or ethnicities, reminiscent of past incidents such as the “Spanish flu.” The term "COVID-19" breaks down as follows: "CO" for corona, "VI" for virus, "D" for disease, and "19" indicating the year it was first recognized.
Despite a decrease in new cases, the global health agency remains vigilant, emphasizing that COVID-19 continues to pose a significant risk. They have urged a prompt acceleration in research efforts to discover effective drugs and vaccines. As of the time this article was published, the statistics were alarming: 60,179 confirmed cases, 1,365 fatalities, and 5,924 recoveries, underscoring the ongoing threat.
Section 1.1: Insilico Medicine's Approach
In the face of this pandemic, efforts to develop a vaccine or treatment are intensifying. Insilico Medicine has announced its pioneering use of AI to expedite the search for molecules that could potentially lead to treatments for COVID-19. Their goal is to compress what would typically take years of research into just weeks.
Starting on January 31, Insilico began analyzing a comprehensive list of potential targets for treating 2019-nCoV provided by the Beijing-based Global Health Drug Discovery Institute. Utilizing 28 different machine learning models, the company was able to identify relevant molecules that could contribute to a cure within just four days.
Subsection 1.1.1: The Role of Machine Learning
The AI system developed by Insilico Medicine successfully pinpointed thousands of new molecules. From this extensive pool, they plan to test the top 100 most promising candidates. The remaining molecular structures will be made publicly available for other researchers interested in exploring these options.
Some of the methodologies employed by Insilico include generative adversarial networks (GANs), which are also used to create deepfakes. These networks were instrumental in generating new molecules with optimal structures for effective drug development. A stringent filtering process was applied to these GAN-produced molecules, allowing Insilico to focus on those with high "drug-like" properties, while eliminating those that were already known.
Section 1.2: Collaborative Efforts in AI Research
Insilico is not alone in this endeavor. Researchers at Michigan State University have also published findings on using machine learning to develop new drug candidates. Additionally, Gilead, a U.S.-based biotech firm, in partnership with a hospital in Beijing, has commenced human trials of an existing antiviral drug, remdesivir, in Wuhan.
Chapter 2: Sharing Knowledge and Inviting Collaboration
In their pursuit of a cure, Insilico has shared their complete research on the platform ResearchGate and their official website. They have also extended an invitation for peer reviews from other researchers to expedite the search for a viable treatment.
The first video titled "How is AI helping to fight the coronavirus COVID-19" discusses the various applications of AI in combating the pandemic. It highlights the role of technology in accelerating research and improving outcomes.
The second video, "Calculating the Cure, AI & Covid-19," delves into specific methodologies used in AI-driven drug discovery and how these innovations are paving the way for potential treatments.