The Centre for Oncology and Immunology (COI), located in Hong Kong Science Park, is a world-class research hub established by The University of Hong Kong (HKU) and funded through the InnoHK Initiative of the HKSAR Government. Comprised of a multidisciplinary group of scientists and clinicians, the Centre is led by Professor Tak Mak, renowned for cloning the human T-cell receptor, and Professor Suet-Yi Leung, a pioneer in gastrointestinal cancer genomics.
COI focuses on gastrointestinal, liver, and blood cancers by integrating immuno-oncology and molecular genetic functional studies. Leveraging advanced technical platforms—including combinatorial CRISPR-based genome editing, single-cell multi-omics, organoid biobanks, medicinal chemistry, and mouse and zebrafish models—the Centre harnesses the power of biotechnology to drive discovery. These local strengths are complemented by prestigious strategic collaborations with the University Health Network (UHN) and the Technical University of Munich (TUM), which contribute world-leading clinical translational expertise.
By collaborating with the HKU Faculty of Medicine (HKUMed) and its Clinical Trial Centre, COI leverages PI-initiated clinical trial capacity while partnering with pharma and biotech leaders for clinical trial-related translational exploration studies. Together, these efforts drive innovative prevention, diagnostics, and treatment regimens, cementing Hong Kong’s position as a premier international oncology nexus.
AI-Powered "Super-Scanner" to Revolutionize Precise Gene Editing
While CRISPR-Cas9 has revolutionized science by "rewriting" DNA, traditional tools often suffer from inefficiency and accidental "off-target" effects. To solve this, Professor Alan Wong, Project Leader at COI and Associate Professor at HKU, led the development of an AI-powered "super-scanner" to achieve surgical precision in gene editing.
The team created two groundbreaking platforms: CombiSEAL and an AI-driven method called TopVIP. CombiSEAL acts as a high-speed scanner, testing and "barcoding" millions of gene-editing variations simultaneously. TopVIP then uses machine learning to predict performance, mapping out optimal designs from less than 1% of experimental data, which saves years of manual labor and costs.
By screening over 64 million variants, the team identified "top performers" that doubled success rates in specific tests while achieving zero accidental "bystander" edits. This patented, award-winning innovation has been featured in world-leading journals, including Nature Methods and Molecular Cell. By providing a cost-effective, unbiased way to build superior gene-editing tools, this platform paves the way for safer, more effective treatments for genetic diseases, representing a major leap forward in the future of therapeutic genome editing and precision medicine.










