Unlocking the hidden: Revealing rare genetic disorders while safeguarding privacy

MedTech & Diagnostics News: Researchers have successfully demonstrated a novel approach to connecting personal records while ensuring privacy protection. The initial implementation of this technique aims to identify cases of rare genetic disorders, with countless potential applications across various sectors.

The significant breakthrough, set to be presented at the prestigious 18th ACM ASIA Conference on Computer and Communications Security in Melbourne, has the potential to revolutionise the field of genetics. Macquarie University researchers have unlocked the means to locate and quantify specific medical conditions while respecting the privacy of individuals. The initial focus is on identifying cases of rare genetic disorders, an endeavour that has remained challenging due to legal and commercial constraints that safeguard personal data.

In a landmark report funded by the Australian government, leading experts from Monash University and several other prestigious institutions emphasised the urgent need for laws to safeguard Australians against life insurance discrimination based on genetic tests. The A-GLIMMER report also concluded that the insurance industry’s self-imposed moratorium aimed at preventing such discrimination falls woefully short of the mark.

Consider the case of a five-year-old boy who shares a mutation in the GPX4 gene with just ten other children worldwide. This rare condition results in skeletal and central nervous system abnormalities, profoundly impacting the lives of those affected. While it is likely that more children with this disorder exist, their privacy has remained guarded within the labyrinthine network of global health and diagnostic databases.

However, imagine a world where medical records associated with rare conditions could be linked and accounted for without compromising privacy. Dr Dinusha Vatsalan and Professor Dali Kaafar, along with the collaboration of Sanath Kumar Ramesh, a software engineer and the boy’s father, have spearheaded this research at Macquarie University’s Cyber Security Hub. Mr Ramesh, also serving as CEO of the OpenTreatments Foundation in Seattle, Washington, initiated and supported this crucial project.

Expressing his excitement, Mr Ramesh stated, “Knowing the prevalence of a particular condition is fundamental to economic assumptions. If a condition was previously estimated to affect 15 patients, and we discover, through comprehensive data gathering from diagnostic testing companies, that there are actually 100 patients, the implications would be immense.”

The economic impact would ripple through various sectors. Companies working on this condition would experience an increase in valuation, while product costing and medical cost calculations by insurance companies would be affected. Diagnostic companies could target the condition more effectively, and the precision of epidemiology would improve dramatically.

Linking and counting scattered data records might appear straightforward, but in reality, it presents numerous challenges, as explained by Professor Kaafar. The absence of a centralised database for rare diseases means records are scattered globally across hundreds of databases. Moreover, the proprietary nature of data presents obstacles to effective collaboration and information sharing.

Technical challenges arise due to the diverse formats, encodings, and storage methods employed by different databases, further complicated by the possibility of individuals being double-counted across multiple databases. As Professor Kaafar highlighted, these obstacles are compounded by the sensitivity of the health data involved.

To address these complexities, Dr Vatsalan and her team employed a cutting-edge technique known as Bloom filter encoding with differential privacy. Through a suite of algorithms, they introduced deliberate noise into the data, effectively obfuscating precise details while enabling the matching and clustering of similar disease records.

To validate the accuracy of their approach, the researchers evaluated their technique using North Carolina voter registration data. The results were remarkable, demonstrating a negligible error rate and ensuring an exceptional level of privacy, even with highly corrupted datasets. This technique surpassed existing methods, showcasing its efficacy and potential for widespread application.

While the research bears relevance in areas such as marketing, cybersecurity, and social media analytics, the team at Macquarie University is particularly passionate about its application to rare diseases.

Professor Kaafar expressed his deep satisfaction, stating, “There is no better feeling for a researcher than witnessing the technology they’ve been developing create a genuine impact and make the world a better place. In this case, the impact is both tangible and paramount.”

Funding for this groundbreaking research was partially provided by the OpenTreatment Foundation, whose commitment to openness and transparency led to the project being completely open-source from its inception. Dr Vatsalan emphasised “The algorithm we implemented is being openly published, aligning with the Foundation’s vision.”

With this technique, Macquarie University researchers have paved the way for a future where rare genetic disorders can be identified, counted, and better understood, all while safeguarding the privacy of individuals. This remarkable achievement holds promise for transforming the landscape of genetic research and ultimately improving the lives of countless individuals affected by these rare conditions.