As use of CRISPR genomic editing technology grows, so do the risks. Revolutionary applications in medicine, food production, and climate are driving an increasing need to accurately detect how and when genomic modifications have occurred—and mitigate their effects on humans.
For those of us who’ve lived with cancer, advances in gene therapies over the past decade offer encouraging possibilities—and hope. Hope for better treatments and better survival rates. Even hope that one day cancer may disappear.
Consider the game-changing genomic editing technology CRISPR, short for Clustered Regularly Interspaced Short Palindromic Repeats. Used in T cell therapy, the ultra-precise capability has demonstrated success in killing certain tumors and preventing cancer relapse.
Beyond cancer, CRISPR aims to create transformative technologies to fight all kinds of human disease. An over $5 billion industry, it’s considered a revolutionary advance for humanity’s health and well-being.
But advances like this bring wide-ranging ethical questions, potential for misuse, and unknowns.
The biggest unknown: the possibility that genetic modifications could lead to mutations now or even generations down the line. How far down the line? Ultimately, all of humanity could reap the advantages—and bear the burdens—of genetic editing.
There’s a fine line between the benefits and the risks. It’s critical to understand if and how genes have been edited. That knowledge will help mitigate the impact on future generations.
MITRE’s CRISPR Detection Project, initiated under our independent R&D program, has expanded our methodology for detecting CRISPR edits. The research team now uses artificial intelligence/machine learning to identify edits much faster and more reliably.
“As we look at eradicating genetic diseases, we’re talking about editing at the embryonic stage,” says MITRE biologist Heath Farris, Ph.D. “You're basically changing every copy of the gene in the entire organism so it doesn't carry that disease gene anymore.”
Farris explains that these types of modifications risk creating biological effects such as destabilizing other genes. Those effects could unwittingly increase the propensity for development of other disease or abnormalities. Genes could also be intentionally edited for malicious purposes, such as creating biothreats.
With so much at stake, insights into gene modifications provide far more than a nice-to-know. Advanced machine learning techniques deliver faster answers, use fewer resources, and help improve outcomes for generations to come.
With this methodology, we've moved the notch forward and shown what can be done. It’s time for others to build on our research and take it in new directions.
From Mice to Men: One Model, Many Scenarios
MITRE originally applied a manual statistical method to determine CRISPR edits. That process enabled limited, direct evaluation of just two genomes. Our research team eventually developed and applied a deep-learning model on 24 CRISPR-edited mouse epigenomes, which essentially describe modifications to a cell’s DNA sequence.
This automated method enables rapid collection of critical data. It also makes detection of edits generalizable, so scientists can apply the method to other research areas.
“A lot of training data goes into building the models,” Farris explains. “A generalizable model that’s been trained deeply on many different things can be applied to many different scenarios.”
Farris, who’s led the project to its near-completion after five years, says, “I love the science. But more than that, I love the technical interactions I’ve had along the way. I do the biology and drive the program, but I tap into knowledge from computer scientists, statisticians, and machine learning experts throughout MITRE and the bio community. I couldn’t do this by myself.”
As MITRE moves toward transitioning this new intellectual property, the project has sparked interest from stakeholders in industry, academia, and government, including the departments of Defense, Agriculture, and Health and Human Services.
Farris and his team hope the methodology can play a small but important part in ensuring responsible use of CRISPR editing—and protecting future generations.
“We’ve moved the notch forward and shown what can be done, which is often MITRE’s role,” he says. “It’s time for others to build on our research and take it in new directions.”
To learn more, contact us at email@example.com.
Join our diverse community of innovators, learners, knowledge-sharers, and risk takers. View our Job Openings.