It’s not easy even for experts to pull useful insights from complex datasets. It’s even harder when the machine learning tools they’re using are only trained to work within a single nation, region, or more specifically, racial population.
Consider the recent popular Netflix documentary Coded Bias. MIT Media Lab researcher Joy Buolamwini discovered that facial recognition does not see dark-skinned faces accurately. She’s pushed for legislation in the US to govern against bias in algorithms. It’s a real problem that many in healthcare are now facing as the use of artificial intelligence in health data analysis becomes more and more prevalent and clinically useful. But the same push for diversity in machine learning reference for health data has been lacking throughout the industry. For this reason, and for the reasons you’ll learn from just one case study below, SOPHiA GENETICS has taken a decentralized and global approach for more than a decade.
A family without answers
A Moroccan couple living in Spain had suffered one of the worst human experiences, not just once, but twice. As they began to expand their new family, they had three pregnancies throughout the span of about one decade. Only one of their children, a healthy baby girl born in 2012, survived. One boy lived for 43 days, and one girl lived only for 20. The immense grief that this brings to parents is immeasurable, and the only thing nearly as horrible than the actual experience of loss is having a lack of answers – never knowing why.
After the death of their infant boy in 2019, the couple worked with regional medical experts in Spain to try and discover the root cause of what seemed to be neurodevelopmental disorders at birth. Researchers started by performing clinical exome sequencing and targeted gene analysis including a Microarray-based Comparative Genomic Hybridization test to look for any abnormalities that would possibly provide more clues. Much to their dismay, the results were inconclusive. The standard of care and testing used at the time were letting them down.
In 2022, they were able to perform more advanced testing after the death of their second daughter. This included TRIO analysis with the SOPHiA DDM platform. TRIO analysis takes clinical exome testing to a deeper level by analyzing data from the infants as well as both parents, giving researchers a better picture of familial, inherited genomic traits. The results finally confirmed that both parents carried a heterozygous frameshift variant. It was inherited by their infants as a homozygous pathogenic variant in the SMPD4 gene, causing neurodevelopmental disorder with microcephaly, arthrogryposis, and structural brain anomalies in the children. While identifying causative variants in recessive genes is challenging through conventional clinical testing, the SOPHiA DDM™ platform’s TRIO analysis enabled the researchers to overcome this limitation.
This family is now better informed about their risk of potential complications in future pregnancies. They’re able to work with genetic counselors to seek guidance regarding family planning that takes the anxiety out of the unknown.
Why diversity matters
In this case, it wasn’t just the type of testing that resulted in a successful discovery. Because SOPHiA GENETICS has worked on a global scale since inception (unlike more common approaches), the SOPHiA DDM platform’s machine learning algorithms have been better trained to analyze the health data of more diverse populations. Meaning, no matter where in the world this family had been trying to seek answers, their clinical researchers would have benefited from more accurate analysis of data that reflected who they are at the core of their genomes.
Every person is unique. We can clearly see this within the more than one million genomic profiles analyzed by our technology. This is why since day one SOPHiA GENETICS has adopted a global approach and placed it at the core of our company’s DNA. We strive to enable better research and health data analytics for all, no matter where they live or where they come from.
If you’d like to learn more about the SOPHiA DDM platform, contact us for a demo or learn more by exploring the many resources of our website.
