Jeffrey Rosenfeld, Ph.D. has extensive experience in genomics and bioinformatics. Over his 15-year career in genomics, he has worked on a wide range of biological and genetic projects including genetic association studies of schizophrenia, genomic testing of embryos for fertility treatments, gene patent litigation, and clinical cancer genome sequencing. For the past five years, he has been an Assistant Professor of Pathology and Laboratory Medicine and the manager of the Biomedical Informatics Shared Resource at the Rutgers Cancer Institute. In this role, he has become very familiar with laboratory testing both for cancer and other genetic diseases. He has set up the computational infrastructure for multiple tumor sequencing panels as well as a clinical data warehouse. He is also a Director at Rosenfeld Consulting.
What were the overarching themes of the conference this year?
Spatial omics was a huge theme this year. Several companies are doing spatial genomics and trying to figure out how to take the market and where to go. The other major category was single cell sequencing, and there were some new developments in this space as well.
Homing in on spatial omics, were there new technological advancements or were the presentations focused on demonstrating the applications of spatial data?
There were technological advancements, as well as data showing the application of spatial technologies. We're still in the very early stages of spatial omics and trying to figure out where we can go with it. The first step is just fine tuning the methods and instrumentation to get consistent data. We are still trying to understand what a “normal” signal is. Even though the field is still developing, we’ve seen it come a long way in the past few years. NanoString’s GeoMx has grown from targeted panels to assessing the whole transcriptome. 10x Genomics is launching compatibility with FFPE samples, which is the standard sample preparation method for cancer biopsies, for Visium. There were also announcements from other spatial players like Resolve Biosciences and Vizgen, which is commercializing MERFISH and I think will be exciting.
As the technology advances, how do you see the market development? Thus far, the single cell market has been dominated by a single player, do you think there will be a similar trend in spatial technology?
We’ll have to see where the data drives today’s major competitors, like NanoString and 10x Genomics. I think they could end up splitting the market between research and clinical, and in both spaces other companies, like Vizgen, will have the potential to develop their platforms. However, I think there is a good chance of them being swallowed up by one of the big guys for their IP, similar to what happened with 10x Genomics and ReadCoor. When considering today’s platforms, Visium is clearly a research machine. It is designed for unbiased research. It’s targeting the same early-stage researchers as 10x Genomics’ Chromium does. The GeoMx, on the other hand, has been compatible with FFPE, the main currency of clinical pathology, from the start and targeted pathologists more naturally. NanoString’s goal appears to be taking this translational research platform and making it into a clinical assay. But with Visium, it's a tool that I have to do a lot of work on my own and it's not set for clinical use right now.
What kind of timeline do you see this CDx development happening over? Are there certain indications or use cases you think will see success first?
I don’t know of a specific cancer type that will have the first CDx, but it would likely map to indications with known targeted therapies. We have a lot of targeted therapies which work decently, but don't work in everybody who has that mutation. There's a lot of mutation resistance that develops against a therapy.. Take a patient with a BRAF mutation for example, you know there are going to be resistance mutations that occur. So instead of just looking at the BRAF mutation, we could look at the expression markers around the tumor to make more informed therapeutic decisions. Once NanoString, or another company, finds the right drug and signature, this is going to become very big, and every lab is going to want to run this analysis. I could see real progress happening on this front within the year. It probably won't be an official diagnostic by that time but people may rely on it. But single cell genomics could also move to this space. More single cell data is being generated right now and this technology has a few years head start on spatial technology.
Are there any other general spatial AGBT takeaways or developments to keep in mind?
There are a bunch of small companies working on spatial and single cell analyses, Vizgen The other movement I saw was the push to sub-cellular level of visualization. NanoString’s is called SMI, which incorporates in-situ sequencing and is expected to launch in 2022.10x Genomics has their own version but we haven’t seen an expected timeline. I think that's interesting to see where that goes but it will take some time.
Pivoting more towards single cell, what were the major takeaways on this front?
10x Genomics announced a little while ago that they were going to offer lower input options. People could run single cell analysis on 1,000 cells and 1,000 genes. Then on the other end you have the Chromium X which can run up to 1 million cells. These offerings are going to bifurcate the market and push things forward, but one of the main questions is how many people are going to need to run 1 million cells. I know some people definitely want it, but we will have to see how this pans out. I think the lower input offering, however, is going to be very big for a couple reasons. One is that in a typical single cell experiment, you don’t need a lot of cells to get your response. If you look at a TSNE plot, a lot of your cells are located in the same area on the plot, I think you can get away with fewer cells. Second is that 10x Genomics stated this will be a third of the cost of the current assays. I think the smaller assays will really drive adoption. It makes costs reasonable, and now I can run a 96 well format assay [e.g., 96 different 1000 cell samples] on one Chromium run. This will really push up the number of samples that can be run, which will be needed to generate the dataset size needed to start pre-clinical translational research.