GB Sciences built a database and AI platform to analyze the healing powers of plants from nine traditional medicine practices.
GB Sciences’ approach to drug discovery combines artificial intelligence and the power of plants to find new treatments based on traditional medicine from around the world. Andrea Small-Howard, its chief science officer and president, said the company is identifying compounds in plants and how the treatments have been used traditionally. “Before computers, drug discovery was done using simplified systems,” she said.
Traditional drug discovery looks for a single active ingredient to treat a single symptom, she said, but most diseases are very complicated. Western medicine’s approach to illness is to treat each symptom individually, which leads to a mix of multiple prescriptions, including medications that are no longer needed and drugs to treat the side effects of other drugs.
“GB Sciences is looking for multi-component drugs based on plants that address a real view of how the body works,” she said. “We are looking at all the different aspects of an illness and trying to deal with them simultaneously.”
GB Sciences uses a human in the loop approach to AI with subject matter experts who help to train the algorithms. Ethnobotanists enter information about the plants and how they are used into the database. Ethnobotany is the study of how a society uses local plants to treat illness and injuries.
SEE: Healthcare is adopting AI much faster since the pandemic began (TechRepublic)
Small-Howard said the company is collecting data from nine well-established traditional medicine systems, such as Traditional Chinese Medicine and Ayurveda. One of the most challenging elements of the work has been creating a database that reflects the wide variety of plants and how they are used.
“We had to upload the data and then make sure all traditional medicine systems were coded in the same way,” she said.
Small-Howard said that the database also had to support multiple queries, such as a search for a plant, a search for a treatment for pain or a search for a plant in a certain region.
“The database construction was the most novel coding we had to do,” she said.
GB Sciences announced in April that it had filed a provisional patent application to protect its proprietary drug discovery platform, which includes a data analytics pipeline, a conceptual framework and machine learning algorithms designed to identify new active ingredients in traditional, plant-based medicines. The company said its goal is to identify complex mixtures of active pharmaceutical ingredients derived from traditional plant-based treatments.
Small-Howard said that one goal is to compile enough evidence to persuade Western doctors to prescribe traditional medicine treatments.
“Areas where these individually evolved systems overlap are likely to predict efficacy,” she said.
The company plans to create a synthetic version of plant compounds that could treat an illness or collection of symptoms, as opposed to using organic material. The company is currently studying kava kava, a plant found on islands in the Pacific Ocean that has the potential to treat anxiety.
Small-Howard said the company has hired people from a variety of sources, including NASA.
“We are looking for people who are happy to adapt what they’ve learned and apply it to our work, basically bright people willing to take a leap on something new,” she said.
The company started with an exclusive focus on cannabis and has developed several potential treatments from that plant for cancer, chronic pain and Parkinson’s. Small-Howard said the company sold the cannabis assets and invested the profits in the drug discovery and validation pipeline.
In 2020, GB Sciences received three patents for treatments for Parkinson’s disease, pain and the anti-inflammatory condition Mast Cell Associated Syndrome. GBS also has two additional U.S. patents and three corresponding patents issued internationally.
A research paper about artificial intelligence in drug discovery and development suggests that AI has the potential to reduce the human workload and make the process faster. Traditional pharmaceutical companies use AI to analyze data sets from millions of compounds.
Another analysis found that about 70% of pharmaceutical companies use AI in some way. The most common task is recruiting and selecting people to participate in clinical trials. Pharmaceutical companies face familiar problems with this new way of working, including data quality and the need to customize existing machine learning tools.