The Goldilocks scenario: Balancing footprint and read depth
Finding the right answers, to the right questions
With continued advancements in next generation sequencing (NGS) technologies, the sky’s the limit! As scientists, we now have the ability to interrogate over 20,000 human genes. When you think about where we were just 20 years ago, it’s amazing to witness the capabilities and opportunities possible today. But, when embarking upon the next scientific questions fueled by NGS, there are still crucial considerations for immuno-oncology applications:
- How many genes do we really need to include? Do I only care about genes within known, canonical pathways, or do I want to discover more novel ones?
- What’s the optimal sequencing depth? Do I care about low abundance mutations, or am I just looking at the predominate mutations present? Is the clonal architecture of the tumor important?
- Is it possible to have the best of both worlds and get high depth sequencing across a broad set of genes?
After careful thought, it likely comes down to the primary objective at hand. An interesting article recently published in Genome Medicine (Garofalo et al. 2016) sparked my interest– the study considered several different approaches in tumor profiling using a small, mid, and large-sized gene panel, as well as a whole exome sequencing assay. These are all very different genomic footprints, and the authors sought to understand which assays might be sufficient for various applications within precision medicine.
Mutational Burden and Neoantigen Load
As expected, the optimal footprint is highly dependent upon the particular research question being asked. For clinical solid tumor profiling, it is logical to focus on those genes with FDA approved therapies or clinical trials available. However, for other objectives like mutational burden detection, a broad assay such as whole exome sequencing (WES), provides complete assessment of the tumor mutational landscape. Yet, this larger assay comes with an associated cost and larger amount of data to mine.
Recent literature reports, as well as data presented at the 2016 ASCO meeting, proposed the ability to extrapolate mutational burden using a large panel. Likewise, the Garafalo study showed that larger panels (~300 genes) might be sufficient for mutational load as it was highly correlated with the WES results from the lung and colon adenocarcinoma samples investigated. Further, the authors show that although neoantigen load also correlated with the WES data, “WES identified a broader spectrum of neoantigens.”
So, is it feasible to find neoantigen load from a panel? The authors go on to mention that a limited number of neoantigens were covered within the panel targets. Further, they discuss that for applications like the development of personalized cancer treatment vaccines, it is really essential to use WES. Since these neoepitopes can reside anywhere, not just the cancer driver genes we’ve all become accustomed to studying, it appears that whole exome sequencing may be the only way to truly detect these elusive targets.
Scientific & clinical objectives need to drive the development of the optimal assays for IO
While the idea of small targeted panels seems attractive in terms of speed, leveraging a broader approach up front may end up beneficial in the long run. For example, one advantage of utilizing a larger genomic footprint is the ability to facilitate studies between cross functional groups. Translational researchers and clinical development arms within an organization may be able to take the same sample, run an all inclusive assay such as WES, and collaboratively leverage that data for their particular purpose. Perhaps one group is focusing on a kinase inhibitor pathway while a neighboring team is concentrating on global gene changes. A single platform can be harmonized between groups.
By really considering the key need at hand, we can find that assay that is just right.