Overview
We spoke with the head of a large academic research center focused on single cell sequencing. We gained some additional insights on Visium, experimental costs and tradeoffs, competition, and penetration. We continue to believe the adoption curve for single cell sequencing, and TXG as the standard platform, is accelerating into a substantial opportunity. As a COVID testing lab, their facility has remained opened as an essential workplace the last several months with modest disruption in scientist access to their labs. We've begun identifying labs new to the space in our analysis of NIH grants which should generate a roster for subsequent outreach. As our contact suggested, "Single cell is here to stay, a component of any lab, like a PCR machine." The most negative comment referred the limitations of Visium compared to the optimism expressed by 10X management.
NIH GRANT UPDATE
Field Notes
Our expert is a molecular cell biologist that started working on single-cell genomics in 2010, and later aligned with the first commercial tools through Fluidigm and its C1 system. His institution is a research community, but he also operates his own research lab. The institute has ~30 faculty, and they support work at the medical school. The lab is funded by both state and federal funding with a primary mandate on single cell research, but also the broader applications in single cell human biology. They use both 10X instruments and consumables often. Research programs include human solid tumors, early human implantation, IPS, and embryonic stem cell differentiation schemes.
The institute managed to maintain research access through COVID-19. Slowed down slightly as graduate students followed university policy and stayed home. Employees and scientists remained open, because they are an essential site for COVID testing and research.
- We did 3 rounds of testing, currently temp screening of employees.
- We had 2 positives out of 350, so debating how to continue testing… plan to continue testing to monitor for outbreak. Wearing masks, etc.
- Able to stay up and running from a research perspective
10X vs Fluidigm, vs Mission, vs Illumina - what are the material differences?
- $TXG: The best because of the throughput primarily, but also the standard for single cell research
- $ILMN: provides a library kit on front- end, but is very low throughput. Co-marketed with Bio-Rad, but never took off.
- Fluidigm: At the time the C1 launched, it was the most efficient technology, utilizing microfluidics and allowing for single cell library preps. The device was capable of capturing 96 individual cells. In parallel, it developed CDNA libraries, added to the cost, and provided great data, but throughput was poor. Regardless, it was better than manual and mouth pipette.
- In 2015, back-to-back droplet-based methods were described in CELL, capturing individual cells. At the same time, molecular biology barcoding was developed for first strand synthesis step. Merging these processes together, scientists were able to do amplification on one sample at a reduced cost per sample. Droplet allowed user to easily do thousands of cells at a time with millions of droplets.
- Fluidigm never solved the throughput problem, because C1 wasn’t droplet, just fluid flow.
- Described as a "platform looking for application."
- C1 was state of the art for single cell through 2015, but as soon as droplets became available, $TXG replaced it.
- Mission Bio: Interesting company with similar droplet-based application. Focus is on looking at genetic alterations, such as analyzing DNA vs. transcriptome. Methods are often targeted and biased. Great for cancer with known repertoire of mutations.
- $TXG: Designs PCR primers to build onto beads that go into droplets, 60-100 mutations per cell. Thinks it works best for targeted genes and specific applications.
- Original single cell analysis on Chromium was pointed at "fresh, frozen tissue." Able to disassociate to single cells, but lost all spatial resolution and where cell sat. Able to generate great view of cell biology.
- Visium, which they adapted from Swedish-based, Spatial Transcriptomics system, is virtually identical, but enable the researcher to spot oligos onto a slide at areas of 6.5mm by 6.5mm. These slides contain about 5.5k individual spots, each containing millions of oligos.
- Run experiment on frozen tissue section, layer over spots. RNA then settles down onto the spots and hybridizes. Then the scientist collects all the oligos, runs the same molecular bio analysis, but now has spatial information in the barcode that is labeled for that 6.5 x 6.5mm.
- Currently, able to do 55 microns spot size, that’s region looking at, any MRNA- large when talking single cells. Depending on tissue, could be 10 or 20, or zero if matrix, not single cell resolution - working on getting spot size smaller
- In his lab, it is protocol to use Visium first or simultaneously with the new- chromium system, helps to interpret spot data in Visium system.
- Example Experiment: Take a fresh bladder tissue, disassociate, characterize cell types, then run a tissue section to overlay on Visium slide for location data.
10X very excited about Visium - what's your take?
- 10X is "pricey" at the moment, definitely extremely attractive in some situations, but not all.
- Less excited for applications in solid tumors, because the cellular structure is “mixed up” vs. a region of brain or structure of bladder where can see pattern.
- Sure they’ll continue developing tech - 10X is great at improving on existing technologies and bringing new ones to market.
- If they can improve the resolution to a single cell location -> lot of potential
- Used it on fresh, frozen, not FFPE - challenging to get good quality RNA out of formalin- need improved chemistry
- Use other methods, highly multiplex antibody staining methods.
- As a replacement for histochemistry
- Don’t have the single cell resolution in the Visium system yet, because the regions on the chip cover multiple cells.
- Regionally within cell, expression would matter - in some instances want sub-cellular, not getting that with Visium now.
- Visium will be useful to some percentage of chromium users, likely as an add on for more sophisticated groups.
- Slide-based - No capital costs, but the slide is very expensive. Can be analyzed on traditional microscope slides that any cell biology lab would have.
- Their chemistry is vital - consumables piece (can’t use other materials)
- Variable cost is roughly 50% $TXG, 50% $ILMN sequencing cost.
- Costs depend on how much sequencing, how many cells, and how many genes are analyzed.
- Time to generate library is approx. 24 hrs - Not a lot of work
- Trying to generate data on 6k-10k cells per 10X Experiment is a challenge
- 50K reads per cell on over 6K cells is about "$1k of of sequencing," but all depends. RNA content virtually parallels cell size. 25-50k reads per cell is sufficient for blood cells, but stem cell are bigger and healthier, require 100k reads.
- Developing pooling strategies, as well as the ability to do targeted sequencing of transcripts. Would pay a little more for targeted sequencing; primers are used to amplify regions. Would spend less overall on sequencing because they would be able to look at targeted genes.
Generic experiment cost is 50/50 10X/Illumina
- $1k for 10k - Sequencing costs have come down. 100k reads per cell would cost close to $2k.
- Academic price for reagents is $1,200 per 10X channel.
- Maximize the cost-effectiveness, but if overloaded, you will get doublets and have to throw that data away. The sweet spot is 6k-10k cells per channel
- Methods to do sample barcoding. If you don’t need 6k, you would barcode 10 samples, and run through one channel.
- Sample prep is an issue as it is hugely important for high quality single cell data
- Limits the number of channels that can be run at one time. You want live, happy cells going through.
- Can’t always do across 8 samples. Difficult to maintain high quality samples to fill an entire chip.
- More often than not, not all 8 channels are run. PBMCs, previously prepared, cryopreserved, and immunologists, can run all 8.
- Sample prep is not expensive, but is considered an "art." - Requires time and attention-to-detail.
Average granted amount - $400-450k - how much goes to 10X and ILMN?
- Materials don't consume much of the award. Average costs are $2K, but need multiples of $2k to have satisfactory data. It all depends on the question, how specific you need to be. Full project may cost a few thousand dollars.
- Has seen grants where almost the entire budget is spent on 10X and Illumina. Types of projects would include human blood profiling and analyzing 500k single cell transcriptomes across many patients.
- Has seen multiple grants where all profiling PBMCs are using 10X and Illumina sequencing, such as Alzheimer's studies and vaccine response for COVID-19.
- NIH has puts out calls with a heightened interest in funding; several RFAs have been focused on single cell projects.
- ENCODE project: Institute within NIH that funds genome sequencing and is looking at the transcript epigenetics structure of cell lines.
- Next phase: All prior funding not cell-type specific or single cell, will now be pushed to functionalize the genome.
- New Projects: Single cell data of human and mouse tissues, tissue mapping centers, generating single cell data to deliver to computational biologists, modelers, etc. will expand purchase of reagents from 10X.
- Pushing whole mapping, single cell resolution for humans and mice.
- Power of Atlases: Compared to baseline, functional studies, knockouts, CRISPR screens will read out single cell. Very functional/helpful. Single cell here to stay as it is a component of any core lab, like a PCR machine.
Is single-cell like NGS, a once in generation opportunity?
- 10X slightly different - Sequencing is one process, DNA is same, feeds into the endpoint for so many experiments. Whereas, one won’t always need to do single cell analysis.
- Best guess: Single cell approach is another piece, tool to probe biology. Not small. Thinks it’s a significant component, more similar to PCR adoption
- Obviously essential piece of molecular lab - Reagents are less expensive vs sequencing
- Will want to do single cell regularly - Will want the tech at hand for when you need it.
Cost and other hindrances to adoption?
- Interpretation is an issue, but software tools are outstanding. Don’t need to be computational biologist.
- Barcoding, pooling will bring down cost.
- New people want to use it, but have no idea how or why. Net new users coming in very high.
- In 2 years, Alzehimer's and Neurology users, went from nothing to now one of biggest users.
- We have 2 Chromiums, just receive 3rd, mainly because of locations - Can have many users using an individual box
All data available upon request. Please reach out to with any inquiries.
