The 2018, 2023 and 2024 Nobel prizes in chemistry have brought us to a present when artificial intelligence has been harnessed to not only figure out the 3D structure of proteins and hence their probable function, but also create proteins tailored for very specific purposes and, in some instances, do so quickly.

Now we’re beginning to see the practical fruit of such amazing achievements applied, in this case, to treat venomous snake bites.

Venomous snake bites occur on all continents except Antarctica, but are most prevalent in India and much of South Asia, Papua New Guinea, the Solomon Islands, Australia, much of sub-Saharan Africa and South America. They are the lowest in Europe.

Despite the fact that Australia has many dangerous venomous snakes, bites are rarely fatal because of public awareness and education, well-stocked antivenom supplies and provision of prompt first-class care.

Unfortunately, that’s not the case in much of Asia and sub-Saharan Africa, where antisera are often not available on-site because of the need to keep them refrigerated and the high cost of the antiserum.

But there’s another problem.

Some families of snakes are especially dangerous because the toxins in the venom are too complex and/or simply fail to trigger a sufficient immune response in horses or pigs used to create antiserums.

That’s the case with the venom from the three-finger family of snakes with over 300 species variants, including the black mamba, coral reef, krait, taipan and cobra snakes, to name the most lethal and infamous.

Enter modern technology — AI to the rescue.

The solution came in four steps.

Step one: determine the precise molecular and 3D structure for each of the major toxins from three-finger snakes using X-ray crystallography.

Step two involved an iterative program called RoseTTAFold diffusion, developed by David Baker, a Nobel Laureate in chemistry in 2024, and colleagues, to create proteins that would precisely fit each of the toxins.

Step three: with those protein models in hand, create matching artificial genes to, in step four, produce in high volume, precise, engineered proteins to neutralize each of the major toxins in three-finger venoms.

There you have it, in four steps.

Of course, getting each step right took time, all except RoseTTAFold. Reportedly, the latter program took less than ten minutes to solve the structure problem for each protein to precisely fit the target toxin.

The beauty of the whole approach is that it can be scaled up to create relatively cheap antisera, tailor-made for each type of toxin that does not require refrigeration.

Does it work? Yes, in laboratory animals. Clinical trials are coming.

In 2024, it was obvious just how powerful these computational programs could be for determining the 3D structure of a protein.

For example, Google’s AlphaFold 2 program solved the 3D structure problem in minutes for individual proteins once the program knew the precise amino acid sequence for each protein and proved to be just as good as X-ray crystallography — a much slower, laborious and much more expensive technique for working out the 3D structure of proteins.

I can’t speak for readers, but I’m stunned with the speed with which new technologies solve some of the most vexing, time-consuming and expensive riddles in biology these days using some version of AI, of which this technology is but one.

Similarly powerful software programs are now indispensable for developing new drugs, designing complex molecular frameworks designed to clean up the environment, capturing carbon dioxide, containing toxic gases and storing hydrogen.

The latter was highlighted in this year’s Nobel Prize in chemistry, awarded to three Laureates, Susumu Kitagawa, Richard Robson and Omar Yaghi, who pioneered the development of metal-organic frameworks, which can capture and store a wide variety of molecules safely.

On a less savoury note, AI is heavily involved in designing military equipment from highly expensive generation five or six stealth aircraft to very cheap and effective drones in Ukraine, Russia and the Middle East, and other programs designed to keep tabs on us with almost every interaction we have with our computer devices.

High-tech AI programs have the potential to substantially improve diagnostic and treatment medicine. On the scale of the universe, operating and analyzing data from the James Webb and the new Rubin telescopes would be impossible without AI. And that’s the case in many other areas in the basic and applied sciences and engineering, which have become heavily dependent on AI.

The six-week series on the 2025 Nobel Prize awards at the Niagara-on-the-Lake Public Library begins on Wednesday, Nov. 5 at 2 p.m. Hope to see you there.

Dr. William Brown is a professor of neurology at McMaster University and co-founder of the InfoHealth series at the Niagara-on-the-Lake Public Library.