0:01
I'm Oriol.
0:01
I'm from Single Cell Discoveries.
0:02
So we are collaborating with Beckman Coulter.
0:05
I am not really working for Beckman Coulter.
0:09
But yeah, so basically I just want to present here the collaboration we've been doing together with 10X as well to really automate high throughput single cell readouts.
0:23
So, but before going into the automation part of it, I would like to introduce a single cell discovery.
0:42
So we are a contract research organisation offering high quality single cell services, but also bulk transcriptomic services.
0:51
We have over 10 years of experience.
0:54
Our founders were performing pioneering work over 10 years ago at the Hubrecht Institute where we spun off in 2018.
1:02
And basically one of also one of the our key aspects of our companies that we have our internal R&D team which developed new assays but continuously improves the currently available ones that we have in our portfolio.
1:14
We have a long tenured service relationship with pharma, biotech and academia and we are a team of 35 PhD level scientists trying to assist all our clients to address their biological questions.
1:31
So the way we work at Single Cell Discoveries is that we are involved from the very beginning of the project.
1:36
So we offer what we say an end-to-end service.
1:39
We are involved in the early consultation calls to try to understand what's a biological question and then define which is the best single cell method or bulk transcriptomic method required for that project.
1:53
Then we can handle any sample type, mainly different sample preparations to reach the library preparation step and then doing the sequencing in house as well with our NovaSeq X Plus.
2:05
And also we have a huge bioinformatics team that can support with the initial data report that we provide with all the services, but also data consultancy projects.
2:16
To go more into the detail of the experiment so far, we have a vast experience with sample types and tissues.
2:27
We've sequenced over 40 organisms so far.
2:30
And between mouse and human, we have already sequenced over 30 tissues.
2:35
We can work with very different formats of samples, either fresh, frozen or fixed and but also a huge variety of sample types from xenografts to organoids and primary tissues and anything in between.
2:50
So going into our portfolio.
2:56
So we have single cell methods, both transcriptomics but also spatial transcriptomics, the mainly the sequencing based ones.
3:08
For the single cell methods, we have our plate based methods that allow you to profile 384 cells in a plate.
3:15
Those are for clients that really cannot obtain huge amount of cells to be profiled at the single cell level.
3:24
But we can offer this solution which is actually super sensitive and gives really nice results.
3:30
So we have a three prime version of it, but also a full length version so they are SORT & VASA-seq.
3:37
Then for higher throughput projects where clients want to profile thousands and millions of cells, we can offer all the flavours of 10X Genomics, but also a scale or parse.
3:48
And then within the bulk transcriptomic services, we have a regular bulk which we are actually really good at it.
3:54
So we can work with very low input samples like hundreds of cells but also we have a very high throughput RNA transcriptomic profiling method which is Discovery-Seq, which has been specifically designed for drug screening purposes.
4:13
Then apart from all these solutions, we also have custom services, tailor made solutions for our clients, for instance biodistribution.
4:22
So we have some clients in the gene therapy space where they are very interested in knowing which cell types are being targeted by their viral particle or viral vector.
4:31
So we can really go into detail and define a custom assay to profile that.
4:36
And also we can offer data consultancy projects to go more into detail.
4:41
More recently also we've become a certified service provider for BioSkryb, which offers whole genome sequencing of DNA on top of RNA of a single cell among other solutions.
4:57
So moving into the automation.
4:59
So it's been an integral part of the company since the very beginning.
5:03
So back in 2018 we already acquired an acoustic liquid handler, an Echo 525 from Beckman Coulter, fully integrated into an access workstation to really help us in the processing of all these plate-based methods that we have in house.
5:18
And later on we decided to start doing our first automation steps for library preparation.
5:24
We acquired an Opentrons, but we quickly realised that was not very scalable and it's a very robust system.
5:30
So eventually in 2023 we acquired the Biomek i7 from Beckman Coulter.
5:40
Before going into how we automated 10X, we'd like to go back into the basis of 10X.
5:46
I'm sure most of you are very familiar with it, but just so that we can understand which part of the method we automated.
5:52
So basically for 10X we use this microfluidics device, the Chromium that basically encapsulates your single cell suspension together with the reagents for the reverse transcription into these oil droplets or gel emulsifications.
6:10
Then you recover this emulsification, these gems after the run and you proceed to do the RT where all the transcripts will be reverse transcribed into cDNA and barcoded so that they can identify which transcripts are coming from which cell.
6:28
Then we move into the library preparation to make this material compatible with a Lumina sequencer.
6:35
And this is actually what we've been automating together with Beckman Coulter.
6:42
So how this collaboration arrived was in a very good time because basically we've been seeing in recent years a huge increase in the number of projects that we're receiving, but also in the size of each of these projects.
6:54
So now every time we have more and more samples from a single project and we always want to be able to provide the data back to our clients within a turn around time of four to six weeks.
7:04
So this means that with this increasing need for throughput, we have to be better doing and running these projects.
7:14
So we were very happy to hear from Beckman and 10X when they reached out for us to be testers of this new kit, which basically automates the library construction of the 10X libraries.
7:28
So this is the workflow overview as I was mentioning, you do the encapsulation with the microfluidic device, you go into the RT, then amplify the cDNA, extract it and do the QCP step.
7:40
And at this stage is when we can move into the automation pipeline.
7:45
So for this you require the Chromium X and then either three prime or five prime solutions from 10X.
7:51
And then for the library construction, the 10X kit that supports this is the new one.
7:57
This library construction kit C automated and the workstation that we're using is the Biomek I7.
8:06
So what happens in the workstation is we go through all the steps of the library preparation from fragmentation and repair and a tailing to the clean ups with SPRI bits, then adapter ligation, post ligation clean up and the sample index PCR.
8:24
If you have an on-deck thermocycler, you can perform the whole workflow within the workstation.
8:31
But in our case, we don't have the on-deck thermocycler.
8:35
So we moved the PCR step off deck.
8:38
And then once the PCR finish, we come back into the end of the workstation to finish the clean up and the final assessment of the libraries.
8:50
So this is what the workstation looks like. On the right side, we have a plate gripper that moves around boxes, but also the plates with the samples.
9:02
And then a multi-channel 8 span that is very useful to prepare all the master mixes that you require during the library preparation.
9:11
And on the left side we have a 96 channel head that allows us to really go into the processing of all the samples in a plate very quickly.
9:23
Then we have some plate cooling systems that keep the reagents at 4°C and including also some shaking capabilities to keep the bits properly re suspended during the cleanups and a wash station that we use to discard many of the reagents.
9:44
So what are the advantages of these automated pipeline?
9:46
So basically in a single run you can prepare 96 libraries.
9:51
This would be 6.5 hours with only one hour of hands on time for the operator.
9:57
Whereas if you were to do a manual run, usually where we feel comfortable for one operator to do up to 32 samples, which would take 8.5 hours with 5.5 hours of hands on time.
10:11
And obviously there's potential human errors and more variability from operator to operator, which is something that the automated run overcomes.
10:20
So going on onto the experimental setup for the validation of this kit.
10:25
So we obtained human PBMCs, 500 cells per sample.
10:30
This was provided by 10X already.
10:32
And what we did is we set up to perform two different runs to validate the inter run of variability.
10:41
But also within each run, we also run the same samples in duplicate to assess variability within the run.
10:48
We tested both samples from the three prime in blue and five prime in green.
10:55
And then we also included non template controls, so basically water to assess potential cross contamination.
11:02
So some of the samples that we had in run one were also present in run 2.
11:06
It's for the comparison and all the numbers indicate the different samples that we had.
11:11
So for reasons of timing, I'm not presenting the all the performances of all the samples.
11:18
But I can tell you basically that the non template controls, there were no reads at all.
11:21
So there was no cross contamination.
11:23
So it's a very clean procedure.
11:25
And then also the duplicates within the same run were identical virtually.
11:33
So here I'm just showing we sequenced these libraries at 30,000 reads per cell.
11:37
And I'm showing here UMAPs for both the three prime and the five prime solutions.
11:42
We also did the match manual run for some of these samples and you can see in UMAP representation without any kind of batch correction on our harmonisation, this is just raw UMAP dimensionality reduction and UMAP projections.
11:56
You can see that there's a really nice overlap for both 3 prime and 5 prime of all the manual samples and the automated runs.
12:03
Then we also ask, OK, are we seeing the same population structure, the same cell types when we do a manual run or an automated one?
12:11
And this is the case indeed with a very high level cell type annotation, as you can see here.
12:18
So then we have the runs at the bottom indicated and then the sample.
12:23
So you can see that for sample one for instance, the manual run and the automated run are identical and then for sample 2 was present in both the first automation run and the second one and they are virtually the same.
12:36
And this is the same for sample 3 then sample four, we don't have a replicate in this case.
12:42
And then for five prime the same repeats even actually the manual ones have more variability.
12:48
So a duplicate of manual had more variability than the automation run for instance.
12:53
And then for sample three, you see automated run one and two.
12:57
Yeah, the proportions are very well maintained.
13:00
So it's a very robust process.
13:05
So in summary, we tested a new automation solution for live preparation for 10X Genomics using this Biomek i7.
13:12
No differences were observed in data quality.
13:15
We obtain the same number of cells and the same population structure when we compare it to a manual run.
13:22
The reproducibility is great even more than the operator manual preparation.
13:29
And yeah, as you I was mentioning, there's a huge increase in throughput, being able to do 96 samples in a very short amount of time but also reducing the hands on time for the operator.
13:42
So if you are in the market looking for partners to run your single cell experiment or spatial, this is some of the reasons why maybe you should pick us.
13:51
So we are really a dedicated team with a lot of experience in the single cell field.
13:56
We always try to help our clients to the best extent.
14:00
We are in very regular communications.
14:03
We are platform agnostic.
14:05
So we always define which is the best approach based on the biological question.
14:11
And yeah, we can also offer bespoke solutions.
14:14
So out-of-the-box kind of request that you wouldn't find in any other commercial kits available.
14:22
And also we provide huge data analysis support for going way deeper into the analysis.
14:30
So yeah, if you are interested, please get in touch and we will be happy to talk to you.
14:35
And yeah, I would like to thank the support of Beckman with this collaboration and obviously 10X for providing the samples and the kits.
14:42
And thanks to all of you for your attention.
14:45
I'll be happy to take questions.
