Unlocking Personalized Healing: The Groundbreaking Interplay of AI, Genetics, and Medicine

Although artificial intelligence does not really transform science itself yet, it transforms its auxiliary tools: data analysis, modeling and interpretation of situations.

This transforms the researcher’s tool (especially for those who are not AI specialists) and this effect is enormous, because the need for researchers to use the help of other disciplines partially disappears (this, by the way, also applies to informatics, because with the help of AI we perform as best as possible the tasks usually classified as social sciences research such as survey analysis).

Returning to this year’s Nobel Prize in Chemistry, we need to delve into computational biology. Since modern computer modeling (which is probably 50 years old, by the way), predicting protein structure from amino acid sequences has been an unsolved problem, largely because of the enormous number of possible protein conformations.

It is theoretically possible for a scientist or all university scientists to count them, but with traditional methods, such as X-ray crystallography or nuclear magnetic resonance, you will not be able to make many combinations, because these studies are incredibly time-consuming and require specific “know-how”.

With the advent of AI (one of the breakthroughs here was platforms such as Google’s DeepMind AlphaFold used by nominees performance), there was a revolution in the field because scientists got the tool.

Such tools have made it possible to predict protein structures with atomic-level accuracy, often surpassing previous simulation results. A faster understanding of biological processes at the molecular level has emerged. Well, the result is a Nobel Prize, because protein structures will drive (or rather, are already driving) advances in everything from drug discovery to synthetic biology.

Medicines – without side effects

True, we should understand that without scientists, researchers who received this bonus for their “know how”, AI would be completely worthless.

But again, let’s not compare AI to people, let’s compare AI to tools. In this “weight class” of scientific tools, AI is perhaps even unrivaled, as it allows relatively easy prediction of 3D protein structures from sequences through huge biological data sets.

By including a person, a scientist, in this chain, it becomes clearer, for example, how to identify potential binding sites of therapeutic compounds and thus accelerate the development of new treatment methods, how to create new proteins that perform specific drug functions. Anyway, the role of such AI is not limited only to biology, such tools change the modeling of various natural sciences (and not only) problems.

First of all, it is the achievements of D. Baker (yesterday’s Nobel Prize laureate) in creating new proteins from scratch that are a big leap in the field of bioengineering. Here, AI is only a tool that helped achieve this, because AI “automated” the exploration of cosmically large protein design spaces, that is, AI allowed to quickly select promising sequences with desirable structural and functional properties, and instead of millions of sequences, it was enough for the scientist to check far fewer of them .

The value of this discovery for science, especially for treatment, is extremely important. This solution will speed up the practical development of drugs, as it will allow much faster modeling of proteins with specific functions, for example, catalyzing chemical reactions or binding to certain molecules with high precision. For example, imagine drugs without side effects. Sound like fiction?

You will always need someone

From the practice of working with AI, I can say that the biggest challenge is the data itself and its quality. Biological systems are inherently complex, characterized by high variability and often incomplete or “noisy” data.

Artificial intelligence models require large amounts of high-quality data to be effective, and biological data are often difficult to obtain or standardize. Furthermore, biological processes span multiple temporal and spatial scales, making model predictions difficult.

When it comes to scientific resources, given the funds, anyone can build a building and fill it with the latest AI computing clusters. However, without the know-how, these investments will be worthless.

And this is the biggest problem, because neither buildings nor machines do science. Of course, we can casually query what ChatGPT clone or AlphaFold itself is installed in that building, but do you understand what you get as a result? And more importantly, are you going to trust ChatGPT or a random AlphaFold issue? Is there a drug with such a protein? Probably not. This will always require a person, a scientist, which is currently less invested in than in AI.

Treatment – according to personal genetic makeup

Optimistically, AI as a more efficient tool could potentially lead to breakthroughs in personalized medicine, where treatments are tailored to a specific person’s genetic makeup, or in synthetic biology, where entirely new organisms can be created.

However, the creation of a specific person or completely new organisms is already a question of scientific ethics, more extraordinary than embryos and cloning. Here we also face a huge risk. Creating synthetic organisms or manipulating genomes can have unintended consequences, and it doesn’t take a sci-fi movie to realize the ecological disruptions or biosecurity threats it can lead to.

But we also have simpler ethical questions. We already have personalized medicine, which is an extremely expensive treatment. If we start to limit the use of AI in the area we are discussing, won’t such drugs become available only to the elites of rich countries who can afford them or end up in the hands of global corporations? After all, this way we will increase inequality in the world even more and even faster.

There is also no guarantee that this technology will not be used, for example, to create new viruses for which the same company will also offer drugs. Sounds like a conspiracy theory, but AI makes it possible in the best possible way.

Let’s not forget that we are moving very fast towards autonomous AI. If we don’t have effective regulation, strict ethical guidelines, oversight, and most importantly a thoughtful and active society, we can get into trouble very quickly. Especially in today’s rapidly digitized fields such as synthetic biology or medicine.

window.fbAsyncInit = function() {
FB.init({
appId: ‘117218911630016’,
version: ‘v2.10’,
status: true,
cookie: false,
xfbml: true
});
};

(function(d, s, id) {
var js, fjs = d.getElementsByTagName(s)[0];
if (d.getElementById(id)) {
return;
}
js = d.createElement(s);
js.id = id;
js.src = “https://connect.facebook.net/lt_LT/sdk.js”;
fjs.parentNode.insertBefore(js, fjs);
}(document, ‘script’, ‘facebook-jssdk’));