0:01
Good morning.
0:01
I'm Patrick Boyd.
0:02
I'm a Field Application Scientist at Scale Biosciences.
0:05
So ordinarily I'd be in the field training people, but today I'm here to talk to you a little bit about everything that Scale Biosciences has to offer.
0:12
As many of you already know, and based off of the conference that we're at, single cell is kind of a major field at the minute.
0:20
But I do want to talk a little bit about why we got into single cell.
0:24
Well, again, I'm sure you all understand single cells really opened up a lot of new information in terms of cell biology.
0:30
We're able to get much more granular information comparing individual cells, looking at the composition of tissues and how those cells interact with each other.
0:36
And obviously we've then been able to apply that same kind of idea to understanding disease as well, looking at disease states, how cells are changing on an individual basis, and how the composition of tissues has also changed at the single cell level as well.
0:49
But one thing that we really care about at scale is kind of giving customers as much flexibility as possible within these kind of assays.
0:56
And one thing we noticed was that a lot of people have been relatively limited in terms of the scale of their experiments.
1:02
And so what I mean by that is that you've been somewhat limited by the number of cells that you can actually process at one time and by the number of samples.
1:09
To this point, they've been, you know, relatively low numbers of cells and relatively low numbers of samples.
1:14
And so part of our mission was to really give you again, as much flexibility and allow you to have that throughput to get much more out of those kinds of experiments.
1:24
But what can you actually get by having more cells in your single cell experiments?
1:28
Well, one thing to think about is better understanding of individual cell types.
1:32
So of course you can find individual cells, but as you can get to that more granular level, get more information about those cells, you can start to break apart differences between those individual cell types.
1:41
So really get a really good understanding of the heterogeneity within individual cell types.
1:46
Additionally, by having kind of increased number of sampling process at one time, that opens up kind of new experiments you may not have been able to look at previously.
1:54
So this could be kind of high throughput drug screening.
1:56
If you've many different types of perturbations, then you're going to start to approach those experiments at the single cell level as well.
2:03
Additionally, we can think about rare cell populations at the same time.
2:07
Ordinarily in some of your experiments, you know, these rare cell populations, there may only be, you know, maybe 1 or 2% of your cell population.
2:14
And so with kind of more standard single cell practises, you may be only picking up 10s or low hundreds of these cell types.
2:20
Whereas if we think about increasing the number of cells within those experiments up into the hundreds of thousands or even the millions as you're going to see, then you're seeing thousands or 10s of thousands of those rare cells.
2:31
And you can actually start to see those within your experiments and they're not being lost within the noise in that single cell data.
2:42
Additionally, as we're kind of going into the area of era of artificial intelligence and machine learning, they require large data sets to be trained on.
2:50
So we really believe that by having these very high throughput assays, again, approaching millions of cells, thousands of samples, it's providing a lot of data to train these on.
2:59
And that'll really help to advance how we're approaching biology using these types of methods.
3:04
But don't just take my word for it.
3:06
I do want to show you a couple of things.
3:08
So in the study on the left, that's combining multiple data sets, 14 different data sets getting out to close to 600,000 individual cells.
3:16
And from there they're able to identify about 6 different cell types they didn't previously identify in individual level experiments.
3:23
Additionally, if we look at the study, just looking at the number of cells within a single cell data set compared to the number of cells that they're identifying, you can see that generally speaking, as you increase the total number of cells, you're going to find more different cell types within your data.
3:36
And so that's really part of what we're trying to do at scale.
3:41
And so kind of the major things I'm going to talk to you today about are QuantumScale single cell RNA products.
3:48
So again, this is millions of cells, thousands of samples for single cell RNA.
3:52
I'm also going to talk to you about methylation today.
3:53
So this is the first single cell methylation kit on the market.
3:57
And then we also do also have a CRISPR screening kit.
3:59
So I'm not going to talk about this specifically today, but if you're doing CRISPR screening, feel free to ask me about that or come and find us at the booth to talk about CRISPR screening.
4:08
Once again, the first thing I want to talk about is our QuantumScale single cell RNA kit.
4:12
If you were at AGBT, you may have, I've already heard a lot about this, but essentially this is our brand new single cell kit.
4:19
And what that looks like is it starts with a fixation up front using a couple of different options.
4:25
This is a methanol based fixation, but we have kind of two options for fixation, one of them being ScalePlex.
4:29
And I'll talk about ScalePlex more later on.
4:31
But essentially what this is a way of having very high throughput fixation.
4:35
So fixing up to close to 100 different sample types at a time, allowing you to have thousands of samples going into your individual single cell experiments.
4:44
Or you can just approach that by fixing individual sample types as well.
4:49
The first step of this is reverse transcription.
4:51
And so individual samples are loaded into this 96 well plate where there's 96 different barcoded RT reactions that happen since they've been barcoded.
5:00
We can then pull all those cells back together, spread them across our quantum barcoding plate.
5:04
And we're going to go into more detail about the quantum barcoding plate on another slide.
5:08
But essentially what happens here is we load all of our cells across that.
5:11
We then add barcoded beads on top of those cells, and we can capture the contents of those cells onto those beads.
5:18
And then finally, we index PCR step.
5:21
And essentially this allows you to actually change the way in which you approach sequencing.
5:26
So we have index PCR that allows you to do either Ultima or Lumina sequencing, giving our customers a lot of flexibility and how they approach that.
5:34
And what this allows us to do overall is to have an output of up to 4 million cells.
5:38
You can process about 9200 samples at the same time.
5:41
And this significantly reduces the amount of time you're dedicating to these.
5:44
So it's about a day and a half workflow and it brings that cost down significantly to close to a penny per cell.
5:53
And so the quantum barcoding plate is kind of the different thing that we've introduced here.
5:57
It's very unique to our workflow.
5:58
And essentially what this is a, again a 96 well plate.
6:02
We refer to each of these wells as a microwell.
6:04
And then within each of these microwells, we have 10s of thousands of microwells.
6:08
And so within each of these microwells, once we add those cells in, you're capturing anywhere from about zero to three individual cells within those wells.
6:16
We then put a barcoded bead on top of those wells and release the contents from those cells and bind that to those well, to those beads and we then carry those beads forward through the remainder of the assay.
6:28
And essentially what this technology allows us to do is to get very high recovery.
6:32
So we recover about 60 to 70% of the sales that are put into that assay allows us to have a very low multiplet rate, so about less than a 4% multiplet rate.
6:41
And then finally to have a very large barcode space.
6:43
We have about 600 million unique barcode combinations that come by having this kind of technology.
6:53
And so again, this is unique to our new QuantumScale RNA assay, but we really see this as a platform for us to continue to build off of.
7:00
And we do plan to continue to increase our throughput as well as bring new modalities to this platform as well.
7:09
And again, I mentioned it's about a day and a half workflow.
7:11
This is much shorter than you would see with other kind of comparable technologies.
7:15
So allowing you to process those samples much faster, having a reduced time at the bench as well.
7:20
One thing I do want you to note is the difference between our small and our large kit.
7:25
So we have 84,000 cell kit and a 4 million cell kit.
7:29
But note there's only a couple of hours difference here.
7:31
So really based off of the number of samples you're processing and the number of cells, there isn't a huge difference in the amount of time it takes to process all those different samples.
7:48
And so in that case, we're talking about just time at the bench, but what really people start to care about is the amount of time they can save across an entire experiment or an entire study.
8:00
So on this left hand side, what we're actually viewing is sample number versus the amount of time it would take you to process that.
8:07
Now, what you can see for our QuantumScale platform, essentially there's no change in the amount of time it takes you to process these samples, whether you're processing, you know, just a single sample or you're going up into the 100 samples it takes in and around, excuse me, sorry, my son brought home sickness from daycare and it's been killing me all week.
8:25
But it still takes around 18 hours to process that, regardless of how many samples you're actually beginning with.
8:30
Whereas again, looking at other technologies, you'd see a linear increase in the amount of time it would take you to process that number of samples.
8:38
Additionally, if we then look at just cost per sample, and in this case we're looking at samples being about 5000 cells, we can see that as you increase the number of samples, we have this significant decrease in cost per sample as well.
8:50
And again, that's because it doesn't really change the workflow.
8:52
You're just changing the number of wells that you use within that workflow.
8:54
And that allows you to significantly decrease cost per sample.
9:00
Again, you're saving a lot of time, you're saving money, but you're not sacrificing any kind of data quality.
9:05
So in this left hand side, we have a PBMC data set where you can, which is about 120,000 individual cells.
9:13
You can have to identify all the major cell types you would expect to see within that PBMC data set with a really good sensitivity across the board.
9:20
Similarly, we have a nuclei mice brain data set here, a little over 1,000 nuclei within this data set, again identifying all the major cell types you would expect with really good sensitivity across the board.
9:31
Again, looking in around 20,000 reads or somewhere around 2000 and 2500 genes being detected within these cells.
9:38
So really good sensitivity within this assay.
9:41
Sorry, excuse me.
9:51
And so as I kind of alluded to earlier on when we were looking at that the time frame required for these experiments, we have multiple different sizes of kit.
9:59
So we have our small and our medium.
10:00
So that's 84 and 168,000 respectively, processing up to about 24 individual samples.
10:06
We then have our large and our extra large which are for 2 million and 4 million respectively and up to 96 individual samples.
10:13
Additionally, you may have noticed we have modular in the middle here.
10:16
And so what modular is again just another way in which we try and provide flexibility for our customers.
10:22
What this allows you to do is actually buy one kit and use that many times over.
10:27
You can use that for, say, processing just one experiment of 2 million, or you can break it down into 12 experiments of 168,000.
10:35
So this is really designed for people who, say, are running a core facility or they need a lot of flexibility in how they approach it.
10:40
Maybe they're getting samples rather sporadically and they need them.
10:43
Again, that flexibility to approach those experiments as they need to.
10:48
And just to give you some more examples of that on this left hand side, we can think about running multiple different experiments in one.
10:54
So again, perhaps you've many different experiments going.
10:56
You've multiple different sample types that you're receiving from different labs.
10:59
You can process all of those simultaneously using this kit.
11:02
You can also take a different approach of just doing them all individually.
11:06
So again, you may just do one experiment 168,000 for each of those individual samples and process those separately.
11:13
Or you may be doing longitudinal studies and you want to process those samples as you receive them.
11:17
You can do that with this kit as well.
11:24
And so again, I want to touch on ScalePlex.
11:25
I mentioned this a little bit earlier on, sorry my throat is killing me, but you can blame my son.
11:35
Apologies.
11:36
Again, with ScalePlex, this allows you to have much higher throughput.
11:39
Essentially what ScalePlex is a 96 well plate.
11:43
Each of these wells contains a unique oligo.
11:45
Essentially what we do is we load our cells into these plates.
11:48
Those oligos bind to these cells.
11:49
Again, these are unique oligos.
11:51
For each of these, we can then combine together all 96 samples into just one sample, which can then be used downstream.
11:59
And so just one example of this is here using 6 ScalePlex plates.
12:03
So that's 576 individual samples.
12:06
There are then pulled together into 6 samples of 6 pooled samples and then load it onto the RT plate.
12:11
We're using two columns per pooled sample.
12:14
And so again, a day and a half workflow to process 576 samples.
12:19
And so again, allowing you to have a lot of very high throughput and a lot of flexibility in how you approach this.
12:26
Just as another example, using the quantum plates as well, we processed 16 frozen brain chunks.
12:33
They were kind of divided into 88 different samples.
12:36
Nuclei were isolated from those and then spread across 8 scale plates.
12:42
Essentially this gives you 768 pseudo samples, which will then process their quantum barcoding to get about 2.8 million nuclei.
12:53
This is divided into two libraries and sequenced individually, and what you'll see is we have one library that we sequenced very deeply.
12:58
It's about 20,000 reads per cell.
13:01
One library was sequenced more shallowly to about 10,000 reads per cell, but you'll see you're still getting very good sensitivity even at this 10,000 reads per cell.
13:08
And so again, this is allowing you to take different approaches to how you sequence these.
13:12
So even if you're going up into the millions of cells, you can see that you can still decrease your sequencing a little bit and get good sensitivity across all these samples.
13:19
And again, what you're seeing here is hundreds of samples, millions of cells with the flexibility at the end and in terms of how you approach your sequencing as well.
13:32
And so whenever we combine the ScalePlex technology with our quantum plates, what that allows you to do is process for the small and the medium up to 2300 individual samples.
13:42
And then for the large and extra-large kit, that's 9200 samples at the same time.
13:46
And again, the actual workflow time doesn't really change for these.
13:49
It's still about a day and a half and up to 4 million cells at a time.
13:52
So again, a lot of flexibility in how you approach this, able to get very good throughput and kind of use this approach for many different types of experiments.
14:02
Again, kind of given you the ultimate flexibility, you can have few samples, a few sales for small pilot studies going all the way up to having a lot of samples with a lot of sales for these very high throughput screening.
14:14
And then to answer a question you hopefully all have is pricing or essentially for the small kit for 84,000, we're under $5000.
14:22
So that's under $0.06 per sale.
14:25
And as we go up to the large kit for 4 million sales, we're about $33,000, which comes to under 1 cent per sale.
14:39
OK.
14:41
And so next what I want to talk to you about is our single cell methylation kit.
14:45
Again, this is the very first single cell methylation kit on the market that allows people to look at methylation across the whole genome and individual cells.
14:55
So how this works is we have a fixation up front is a formaldehyde based fixation allowing you to store for up to four weeks.
15:02
We then spread those fixed nuclei across 3 tagmentation plates where we introduce 288 unique barcodes.
15:09
We then sort into plates and then do bisulfide conversion and follow that up with an index PCR reaction afterwards.
15:18
What we can see is across many different experiments with both human and mice with good correlation between total number of reads and total unique reads as well.
15:27
And again, we see the same linear relationship between median unique reads and median CG coverage between both human and mouse samples across many different studies.
15:35
So a very robust experiment here.
15:38
Additionally, we wanted to demonstrate this with more challenging.
15:41
So we've looked at a number of different PBMC samples, kidney samples as well as OCT embedded tissue and again continuing to see robust performance across these different sample types.
15:55
In this experiment, we're looking at PBMCs.
15:57
So we have about 2000 PBMCs.
15:59
We can see a really good cut off between true nuclei and background.
16:02
So again, very clean data here and at about just over about two and a half million reads per cell, we're doing very good both CG and CH coverage.
16:11
We're about 78% CG methylation and a little over 0.3% CH methylation in this case,.
16:16
Again, you can identify all the major PBMC cell types you would expect to see within this.
16:24
And again, you can look at these UMAPs with differential methylation across this as well and see the different methylation patterns present across these different PBMC cell types.
16:35
In this case, we're looking at a slightly more challenging experiments, so this is glioma.
16:39
We're looking at about 8000 nuclei.
16:40
Again, very clean data with this knee plot able to call between true nuclei and background and again identifying all of the major cell types we would expect to see as well as tissue specific cell types such as cancer stem cells present within these samples.
16:57
Additionally, again we can look at CG and CH methylation across these sample types and see differential methylation across samples.
17:04
So we see higher methylation in these groups versus these groups and again, lower CH methylation within these populations here.
17:14
And finally, we can also cluster based off of CH methylation.
17:17
Again, see differential CH methylation across different cell types here.
17:25
Finally, I do also want to note that this is compatible with the Twist Human Methylome Panel.
17:30
Essentially what this allows you to do is enrich for promoter regions and reduce your sequencing requirements.
17:34
So one thing is that for typical experiments, right, about 2 million reads per cell for whole genome coverage, whereas with the Twist Human Methylome Panel, we can bring that down to about half a million and still have very good enrichment of those promoter regions and identify all of its major cell types that you would still be looking for.
17:48
So again, providing a lot of flexibility how our customers can approach their experiments.
17:55
And so in summary, so there's you know, unbiased whole genome single cell methylation across both CG and CH sites, flexible throughput.
18:05
So I know from what 4000 to 18,000 nuclei depending on the kit size, you can have that target enrichment for the Twist Methylome Panel, looking up specific perimeter regions, streamlined workflow.
18:17
So we've made a lot of improvements to this to make this as easy as possible.
18:20
So no equipment is required.
18:23
And then again robust performance across many different cell types.
18:25
And we again we continue to improve the [unclear], reagents required, and things like that.
18:32
So just to summarise, we saw about QuantumScale single cell RNA up to 4 million cells, thousands of samples at the same time, totally instrument free workflow, first single cell methylation up to 16,000 nuclei, again instrument free and the very first single cell methylation kit to the market.
18:50
I just want to highlight we do have in house services.
18:53
We do offer a lot of one-on-one support as well.
18:55
So every single customer that buys a kit, we go on site, and we train them, we support them the entire way through those experiments.
19:01
I'd also encourage you to check out Scale University where you can find a lot of information on getting started with Single cell and about our products as well.
19:07
And then find our support website.
19:08
You can find all of our protocols as well as sample data sets if you want to explore those as well.
19:13
So finally, thank you for being here today.
19:15
Appreciate you taking the time to listen to my talk.