Catharine Aquino’s Post

There’s a lot of hype around Roche SBX pre-AGBT, largely driven by social media. After learning more, here are my thoughts. Take note that I’m approaching this as a genomics technologist focused on operations and implementation. The key questions I asked (and my SB would surely want to know 😅): What value does it bring? What unmet needs does it address? How stable will it be? Do we really need an instrument with these specs? Which research projects will benefit? Which workflows will it replace? SBX’s biggest strength is scalability and speed. The dream has been a machine functioning as both a small sequencer and a high-output system. That said, SBX is not portable—it’s two large boxes. Generating billions of reads quickly sounds great, but NovaSeq X is also fast, and we’re already struggling with data transfer and demultiplexing (DMX) taking longer than the run. Speed is useless if IT can’t handle it. The tech seems over-engineered (impressive but still...), making me skeptical about real-world robustness. Broad got good data fast, but the Broad is not an average lab. Illumina’s success came from reliability. I saw firsthand how the HiSeq 2000 outperformed SoLiD, 454, and Ion Torrent because it was simply easier to use. Sequencing platforms win or lose in the lab, not on paper. The silent determinant of an instrument’s success is the wet lab experience. No matter how good the chemistry sounds, SBX sounds complicated. I suspect SBX will struggle with complex samples beyond cDNA, human WGS, and 10X libraries. This reminds me of PacBio—it works well but has stringent sample requirements. Is 1,000+ bases long enough? Better than PE150, but NovaSeq X can now do PE300. What can SBX do that we can’t already do? We still need long-read sequencers. Mid-length reads don’t change the game—they’re just another option. For new applications, Constellation is more exciting than SBX, with potential in structural variants, methylation and DNA interactions. Based on AGBT trends, my fearless forecast is SBX will take at least two years to hit the market. Would I consider it if it weren’t from Roche? Probably not. But Roche has the resources to refine it, so I’m watching. If I had to buy an instrument today—factoring in all the roadmaps I’ve seen this week and assuming limited resources—I’d still buy a NovaSeq X 😅. PS. Waiting for my flight back home. Nothing else to do other than reflect about the past week #agbt25