AI Advancements in Biotechnology

By Newnew Zhang ‘28, for Helix Magazine

Recently, the field of biotechnology has been revolutionized by Artificial Intelligence (AI), with it now being able to powerfully assist the process of therapeutic development, also known as drug discovery. Before the rise of AI, drug discovery took scientists years to conduct, with plenty of obstacles such as high attrition rates and the need for immense investment. In many cases, scientists took more than a decade to discover new drugs, only to fail clinical trials and lose approval. However, with the assistance of recent advancements in AI, scientists can now accelerate the development of new drugs by using the technology to support them in different stages of research and to discover connections between large amounts of data efficiently. 

Most scientists begin the process of therapeutic development by identifying a disease that requires treatment. Researchers collect a substantial amount of data in order to characterize the nature of an illness. Rare diseases are often difficult to diagnose and very time-consuming as doctors must rely on heavy manual interpretation and observation to identify them. However, the introduction of AI has vastly streamlined disease identification. According to Science Direct, “AI algorithms have demonstrated promising outcomes in the diagnosis, prediction, and management of non-communicable diseases, including diabetes, Alzheimer’s disease, and cancer.” The application of AI can strongly facilitate the identification of infectious diseases, as it can quickly examine vast and complex datasets from resources such as social media, electronic health records, and news reports. 

The usage of AI has also enabled researchers to create reliable diagnostic systems for various diseases, allowing them to identify potential targets for new therapeutics. Target identification, a crucial step in developing therapeutics, often takes up enormous amounts of time in traditional methods. This step involves locating specific proteins, genes, and molecules that are likely drivers of the disease and could be influenced by a drug. According to the DrugBank research team, “One of the most notable applications of AI in target identification is its ability to analyze omics data (genomic, transcriptomic, proteomic, and metabolomics) to identify biomarkers that can serve as therapeutic targets.” AI’s proficiency in processing and pinpointing complex relationships between proteins, genes, and diseases allow for the identification of promising targets that humans may have overlooked, enabling scientists to narrow down targets and accelerate the development process. 

After identifying the target, scientists begin creating molecules that can effectively interact with the disease’s selected target. Conventionally, this requires immense laboratory work to ensure that the molecule will bind perfectly to its target. However, as Science Direct stated, “Through molecular generation techniques, AI facilitates the creation of novel drug molecules, predicting their properties and activities, while virtual screening (VS) optimizes drug candidates.” Advancements in AI have made this process much easier, as through its generative models that can digitally invent extremely precise 3D molecules. AI also makes it possible to rapidly test millions of molecules against a specific target with an accuracy of seventy-five percent during virtual screening. This selection gives scientists a narrow set of the most promising molecules to physically synthesize in the lab, thereby reducing the cost and volume of the experiments that scientists would need to conduct. 

Traditional methods of drug discovery take about three to six years, especially as the proposed drugs must undergo preclinical testing, during which they are tested in cell structures and animal models to assess its safety. This stage requires multiple rounds of chemical modifications and lab experiments absent the help of AI. However, because AI can predict the behavior of a molecule in the body, including its toxicity, it will suggest modifications before lab testing, saving a significant amount of time and resources. 

One example of AI’s transformative impact on biotechnology is the case of Insilico Medicine. Founded by Alex Zhavoronkon and Alex Aliper in 2014, the company was established to use AI for efficient and rapid drug discovery. Insilico utilized its two AI platforms, PandaOmics and Chemistry42, to identify novel disease targets and digitally design new molecules to act as potential drugs from scratch. According to Insilico, “we performed all the required human patient cell, tissue, and animal validation experiments to claim a first-in-class preclinical candidate for a novel pan-fibrotic target, currently in preparation for clinical development.” This statement demonstrates how AI can now be used to develop a completely new type of drug that can treat several diseases and pass early tests all the way to the preclinical stage. Even though traditional therapeutic development costs billions of dollars and takes up anywhere between a year to fifteen years, Insillico achieved this milestone using approximately two-point-six million dollars in a span of just eighteen months. This breakthrough marks a turning point in the integration of AI into biotechnology and reveals the notable advancements of AI.