A newly developed method combining low-cost, deep whole-genome sequencing (WGS) with duplex error correction has achieved unprecedented sensitivity in detecting cancer DNA from blood samples. The approach – developed by researchers at Weill Cornell Medicine and the New York Genome Center (NYGC), USA, and based on Ultima Genomics’ sequencing platform – may allow for plasma-only monitoring of cancer recurrence, treatment response, and potentially early detection, without requiring matched tumor samples.
We caught up with lead author Alexandre Pellan Cheng (Associate Professor, École de technologie supérieure, Canada) and corresponding author Dan A. Landau (Professor and Core Member, NYGC) to find out more.
What was your main inspiration for this study?
Dan Landau: Back in 2019 we were trying to get our MRDetect paper published and had to counter a lot of resistance in the field that viewed the cost of WGS as prohibitive to clinical application. We made the argument that genome-wide mutational integration in cfDNA could have a transformative impact as an orthogonal approach to the prevailing paradigm of deep targeted sequencing. Specifically, because plasma WGS is not as limited by input material as deep targeted sequencing. However, WGS cost at that time was well beyond what would make sense for clinical application. We argued that sequencing cost would continue to decline – and even snuck in an in silico assessment showing that deeper sequencing (120X) could allow us to detect even one part per million with simple WGS. However, the decline of sequencing cost stalled, making the work with Ultima (still deep in stealth) an exciting prospect – as a way to help disrupt the sequencing market that was dominated by one provider.
Tell us more about why you decided to leverage Ultima Genomics's platform?
Alexandre Cheng: As Dan mentioned, Ultima was still in stealth mode when they approached us to work together; and, as technology developers, it was an enticing prospect. The appeal of low-cost sequencing is obvious, but we were also very curious about the underlying technology. Ultima's flow-based chemistry is very different from other sequencing methods, and we wanted to explore it in depth. Under certain circumstances, we thought that the chemistry itself could be best-in-class. We benchmarked Ultima sequencing to standard sequencing methods and found it to perform quite similarly in terms of ctDNA detection.
Did any of the findings surprise you? Any big "eureka moments”?
Landau: A pivotal moment was the confirmation that the sequencing accuracy is strong – bringing to life our prediction that 100-120X WGS can allow ppm sensitivity. We then thought, "what’s next?” What would be a non-incremental of leveraging the markedly lower sequencing cost? However, although deeper (120X vs 30X) WGS can increase sensitivity, it's still only an incremental step forward. But then realized that the lower sequencing cost can allow us to be more ambitious – to implement for the first time molecular error correction at the genome scale.
The success of this work is really exciting because it paves the way for removing the dependence on matched tumor tissue in MRD testing and potentially opens the possibility of plasma WGS for early detection or screening.
Cheng: Our goal was to build a ctDNA detection platform without the traditional requirement for tumor-tissue sequencing, and we were thrilled to see that pan out. What was surprising, however, was our platform's ability to potentially detect the genotoxic effect of chemotherapy in normal cells. We touch on this a little bit in the paper, where we show that patients with virtually undetectable ctDNA still carry a substantial amount of chemotherapy-derived mutations. This suggests that normal cells carry these mutations and suggests that our platform can do more than detect cancer – it can also monitor for off target effects of treatment.
What are the main barriers to implementing low-cost genomic technologies into standard clinical practice?
Landau: There are a few barriers. The broader field is still digesting the new ecosystem that has low cost sequencing and the transition from diagnostics that are anchored in targeted panel takes some thought and investment. That is both on the production side as well as in translating some of the machine learning implementations into clinical grade diagnostics. We are seeing a lot of enthusiasm from large industry diagnostics groups so I hope we can accelerate this process as a community.
Cheng: We need more machines in the field, and more scientists using them and publishing their findings. If you look back at early days of next generation sequencing, a lot of studies that focused on characterising the technology were published. I think we need another wave of tech evaluation papers so we can design the best clinical tests afterwards.
What does the development of low-cost genome sequencing platforms mean for how we think about the legacy of the Human Genome Project?
Cheng: I think the development of low-cost sequencing platforms significantly expands the legacy of the Human Genome Project. I like to think of low-cost sequencing as the gateway to our internet era. If costs drop by another order of magnitude or two, sequencing could be as ubiquitous as an internet connection, and, similarly to your Internet plan, it might make sense to have an unlimited sequencing plan. At that point, the Human Genome Project would be the world wide web of sequencing, and what a legacy that would be.
Do you have any future plans for this work?
Landau: Yes, we have developments across multiple fronts: methods to enhance the efficiency of duplex sequencing; methods that integrate analytic and molecular noise reduction in plasma WGS; and clinical applications enabling the detection of early-stage and low-burden disease in the absence of matched tumors.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
