Can Genomics Redefine the Healthcare System Without Compromising It?
With the support of AI, genomic data science can generate more personalized, actionable information that improves health outcomes — but only if we leverage it correctly.
Our DNA carries a host of genetic information that can help us improve our health and advance society as a whole. Researchers have spent decades studying DNA through the field of genomics, but how we access this data has largely remained a mystery — until now.
Significant advances in the data science arena are empowering researchers with the complex statistical and computational methods they need to crack DNA sequences once believed to be impenetrable. Thanks to these advances, an estimated two to 40 exabytes of genetic data will be unlocked over the course of the next decade.
To help us better understand and leverage the data generated through genomic data science, researchers across the globe are turning to Artificial Intelligence (AI). But while the union of genomics, data science, and AI can prove deeply rewarding, it also carries inherent risks.
Paving the Way for Greater Personalization and Precision in Healthcare
As humans, there is only so much information we can process at any given moment. Think of being asked to identify how many times the word “when” is used in a stack of 500 articles. It will take you hours to complete this task, and the margin for error is high.
Thanks to AI, many manual tasks once performed by humans are now automated. AI-driven solutions can transcribe our voice commands into texts, staff customer help desks, guide our gps systems, and so much more. But perhaps some of the most exciting AI advances have occurred in healthcare, where researchers are leveraging AI to decode our genetic data and identify patterns that could signify type two diabetes, cancer, coronary artery disease, and more.
With these insights, healthcare is making a much-needed shift towards personalization. When our doctors know what conditions we are genetically predisposed to, they can help us take a more proactive, customized approach to monitoring and treating them.
Take actor Angelina Jolie for example. After discovering she had the BRCA1 gene mutation, which can be a sign of breast cancer, Jolie had bilateral mastectomies to minimize her chances of developing the deadly disease. Although controversial, women across the globe followed suit as part of the “Angelina Effect.”
This concept is known as precision medicine, and it’s gaining traction across the globe. But while genomic data science and precision medicine have come a long way, they’re still experimental fields that we must approach with caution.
Eliminating Biases in Genomics and AI Decision Making
It’s hard to imagine how technology can be biased — it’s not a living, breathing human, so how can it discriminate? But studies have proven that biases not only exist in AI, they’re commonplace, and they create a ripple effect of discrimination within the healthcare system.
To understand this concept, we must look at the inputs AI-driven platforms are examining, not the technology itself. If you don’t include genes or specimens from a diverse set of individuals, then the data you generate will be biased towards a very specific set of the population. This data will be used to build models, set thresholds, and define processes for all individuals — even those that weren’t included in the sample set.
Since an estimated 78% of participants in genome-wide association studies (GWAS) are individuals of European descent, the vast majority of research used to identify genetic risk variants fails to accurately represent all populations.
The Path Forward for Genomic Data Science and AI
Like all experimental fields, genomic data science has its risks, but it also has its rewards. When we take steps to amass and analyze data from all populations, we’re demonstrating our commitment to a more inclusive healthcare system.
Initiatives like the National Institutes’ of Health All of Us campaign are leading the way and paving the path for greater equality in healthcare. Instead of focusing on one population, All of Us aims to build, “a diverse database that can inform thousands of studies on a variety of health conditions.”
The result is a healthier society with fewer barriers to treatment — for everyone.