Artificial Intelligence is starting to change how scientists understand and develop treatments for ataxia-telangiectasia. In short, AI allows computers to sift through huge amounts of data, find patterns people might miss, and help scientists make smarter decisions faster. While we’re still in the early days, researchers are already using AI to open up new paths toward better diagnosis, faster drug discovery, and smarter clinical trials for A-T. Here are some of the ways AI is being used—or could soon be used—to help families living with A-T:
Understanding What’s Going Wrong in A-T
To fix a problem, we first need to understand it. With A-T, the ATM gene is disrupted in every cell, but scientists still don’t fully understand why some types of brain and immune cells are affected more than others. AI tools are now being used to study massive datasets from human brain tissue and blood samples. These tools can find key changes in specific cell types, like the brain’s Purkinje neurons or immune cells called microglia, that help explain the early damage seen in A-T. This kind of analysis helps researchers zero in on exactly which biological pathways are going wrong, which in turn can point to new ways to treat the disease.
Better Classification of Genetic Variants
There are thousands of different ATM gene mutations, and not all of them have the same effect. Some are clearly harmful, some seem harmless—and many fall into a gray area. AI can help here, too. New AI models can predict whether a particular mutation is likely to be harmful based on how the DNA sequence looks, how evolution has treated similar changes in other species, and what we know about protein structure. This helps doctors and genetic counselors make clearer diagnoses and allows researchers to better group patients for clinical trials.
Finding New Drugs (and Repurposing Old Ones)
One of the fastest ways to help patients is to find existing drugs that might work for A-T. AI is being used to do just that. For example, researchers can now take pictures of A-T cells under a microscope after they’re treated with thousands of different drugs. Then, using machine learning (a form of AI), they look for changes that make the sick cells look more like healthy ones. The computer helps rank which drugs have the biggest positive effect. Several promising compounds, some of them already FDA-approved for other diseases, have been identified this way.
Designing Brand-New Medicines Using AI
When existing drugs aren’t enough, researchers try to design new ones from scratch. That used to take years. Now, thanks to AI, scientists can generate entirely new molecules on a computer—ones that fit into the target protein like a key in a lock—and predict which ones are likely to work. Some of these AI-designed molecules can be made in a lab and tested in cells or animals within weeks instead of months or years. This approach is now being used in other diseases to design new drugs that target the same stress-response systems affected in A-T.
Creating Better Antisense and mRNA Therapies
Another exciting approach for A-T is to help the body make working ATM protein again using synthetic instructions, like antisense oligonucleotides (ASOs) or messenger RNA (mRNA). But designing these therapies is tricky. They have to be strong, safe, and able to get into the right cells in the brain. AI can help by quickly testing millions of potential sequences on a computer and predicting which ones are most likely to succeed. This kind of optimization helps researchers focus only on the most promising candidates, saving time and money.
Tracking the Disease with Better Biomarkers
To test whether a treatment is working, we need good ways to measure changes in the disease. These measurements are called “biomarkers.” AI is helping identify two kinds of biomarkers for A-T:
- Digital biomarkers: AI can analyze speech recordings, walking patterns, or smartphone video to measure how someone’s condition is changing over time without needing a hospital visit.
- Molecular biomarkers: AI can also analyze proteins and molecules in blood samples to detect early signs of brain inflammation or oxidative stress, both major problems in A-T. These blood-based markers could be used in future clinical trials to track how well a drug is working.
Designing Smarter Clinical Trials
AI is also helping companies and researchers run clinical trials more efficiently. By learning from past patient data, AI systems can recommend better ways to group patients, select trial sites, and adjust a trial as it goes, saving time and reducing the number of participants needed. After a drug is approved, AI can also help detect any rare safety issues more quickly by scanning health records and medical reports.
Challenges and Opportunities
Of course, there are still hurdles to overcome. AI systems need a lot of data to work well, and A-T is a rare disease. That means researchers need to collaborate, share information, and build datasets that are diverse and large enough to train accurate AI models. It’s also important that the results are understandable and trusted by doctors and families. That’s why researchers are working on ways to make AI predictions more transparent, so we know why a certain drug, mutation, or treatment approach is being recommended.
Looking Ahead
AI is not a cure, but it’s becoming an increasingly powerful tool to speed up progress across the entire research and drug development process. In just the past few years, AI has gone from a futuristic concept to a practical partner in fighting rare diseases like A-T. If we continue investing in data, collaboration, and smart science, AI could help deliver meaningful treatments for A-T sooner than anyone thought possible. Families, donors, and scientists each have a role to play in this progress, and AI is giving us new hope that working together, we can beat this disease.
– Brad Margus, A-T Children’s Project Founder, Volunteer Board Chair and A-T Dad