How Can AI Help Discover ‘New’ Drugs Against COVID-19 In Brazil?

Brazilian researchers have topped a bioinformatics competition looking to find already-approved drugs that could function as treatments for Covid-19.

The need for new treatments has not waned: the virus was still the 10th highest cause of death in the US in 2023 according to figures from the Centers for Disease Control and Prevention.

Karina dos Santos Machado , an associate professor at Universidade Federal do Rio Grande in Brazil, led the team that topped a competition help by Canadian nonprofit biotech organization Conscience finding the most molecules that may lead to a new treatment effective against all coronaviruses.

Machado explains that drug discovery can be expensive and long , so re-purposing existing medicines for new uses can streamline this process.

"We tested in the computer only already approved molecules, such as those approved by regulatory agencies as the FDA," she says, "In drug re-purposing, algorithm-driven drug discovery is highly promising, especially in situations with limited resources for testing or where there is a gap between the academia and the pharmaceutical industry as happens in Brazil," she says.

Machado explains that when scientists search for new drugs, they need to understand how a potential medicine interacts with a target, typically a protein, using computer simulations to predict how well a drug candidate fits to a target protein, much like fitting a key into a lock.

"All these in-silico (computer-simulated) strategies supported with artificial intelligence have the potential to make the process of drug discovery easier and quicker in Brazil," she says, adding that these techniques are already being developed and applied to investigate new therapeutic compounds, including those derived from plants and animals of Brazil's rich biodiversity.

Machado says the full integration with these in-silico approaches into Brazil's pharmaceutical industry is still in its early stages.

"Bridging this gap would be a significant step forward in accelerating drug discovery for neglected diseases so present in our country," she says, "I really hope that our victory in the Cache Challenge #2 can help in this sense, highlight the potential of the research done in universities for the pharmaceutical industry."

Growing up in Brazil

Machado grew up in Bage, a southern Brazilian city on the border with Uruguay and says she was drawn to numbers and logical challenges, as well as calculators, typewriters and video games.

"When I was finishing my elementary/middle school I asked my parents to enroll me in a technical high school focused in informatics; I had seen computers on TV but I had never used one." she says adding that she received her first computer on her 15th birthday, a significant feat in Brazil at the time.

Machado explains that her passion for computers led her to study computer engineering at a public university near Bage, a Master's degree in computer science focused on bioinformatics (on a scholarship) and a Phd focused on structural bioinformatics.

"Since 2005 I have been working in this field and have always dreamed of discovering a new drug for a disease," she says.

Machado explains that in Brazil, much of the scientific research comes from public universities, where researchers must balance teaching classes, administrative work, supervising students, and research, all on a shoe-string budget.

"The distinct environmental conditions, including climate, biodiversity, and ecosystems, require specific approaches that may not be addressed by studies conducted in the Global North," she says, "Additionally, the unique genetic diversity found in local populations offers invaluable insights that can contribute to global scientific knowledge."

Machado explains that in Brazil, public universities closely integrate their work with the community, frequently working on problems that affect the local population such as tropical diseases, climate change and its local consequences, usage of biodiversity as raw material and many other areas.

"I believe that this perspective of collaboration, of helping each other, dedication and love for science make our perspectives unique when compared with the Global North."

Machine Learning In Coronavirus Data

During the height of the Covid-19 pandemic, a physician-turned-entrepreneur raised in Kashmir used big data and machine learning to help detect useful patterns in the tsunami of public health data generated world-wide by the COVID-19 crisis.

Junaid Nabi , a public health researcher now working for the World Health Organization, says his experiences with the health system in the developing world drove this past project during the pandemic.

"Growing up in Kashmir, a society marred with social, economic, and healthcare disparities, I was exposed to the inherent inequities in my community at an early age," he said, "During the final years of my training, I had an opportunity to work with some non-profit organizations, especially the rescue teams during the Savar building collapse in Dhaka, Bangladesh."

Nabi, who is also an Aspen New Voices Fellow, worked with colleagues at Harvard Medical School and Harvard School of Public Health to develop digital tools that harness big data and machine learning to rapidly evaluate patterns in the data pouring in from clinical research.

"I believe machine learning has an important role in COVID-19," he says.