0:00
Hi everyone.
0:01
Yeah, my name is Riccardo.
0:02
I'm one of the co-founders at Alithea.
0:05
Today I'm going to be talking about a relatively new technology that we have launched since now a couple of years and that we have seen as really been used widely by our user base.
0:18
A couple of things about Alithea.
0:19
Alithea is an EPFL spin off.
0:21
So we are a Swiss company.
0:24
We're now a Swiss American company.
0:26
Actually we have two sites, one in Switzerland and one in Maryland in the US where we actually do distribution and our services as well.
0:35
So we have these two locations around the world.
0:39
Alithea is focused on the development of new tools for the generation of transcriptomic data.
0:48
We are one of the companies that you may say is helping to improve the way we develop new drugs, the way we actually develop and screen new compounds.
0:58
And we're taking the approach of the point of view of data generation.
1:02
Of course, there are other dimensions like improving the model systems that we use and the way we analyse the data.
1:10
We integrate the data, and the overall overarching goal is to really be able to take decisions earlier in the pipeline to be able to actually understand the biology of the system and what the chemicals that we're studying do to the system earlier than that was possible before.
1:25
Conceptually, what we typically look at is way to transform the typical V shape drug development funnel into the T shape in which we basically take decisions earlier on to really save money and time down the road.
1:40
As I was mentioning, Alithea is a focus on one of these aspects which is actually data generation and specifically transcriptomics.
1:48
So we are hyper focused on high throughput transcriptomics and the technology that I wanted to present today is really in a way summarising and fitting well in this dimension.
2:00
DRUG-Seq is a technology that we developed to be we offering the best of both worlds.
2:07
The 2 worlds being: 1. Targeted assays in which you actually have to predetermine what you're measuring, like microarrays, targeted assays.
2:17
And the idea of these ones was that the advantage was high throughput and very low cost.
2:23
But then it of course offers a disadvantage, which is the fact that you have to know a priori a little bit about the biology of the model that you're studying.
2:33
And the assay is per definition in this case biassed.
2:40
The other side is the RNA sequencing side.
2:43
RNA-Seq is of course a very interesting and exciting technology because it allows you to actually measure everything that is expressed that is RNA in an unbiased manner.
2:56
That current implementations are of course very expensive and low throughput.
3:00
And the idea of DRUG-Seq that especially the DRUG-Seq version that we have developed at Alithea is to really offer a solution to these trade-offs and offer a solution that is extremely high throughput, actually higher throughput than cell painting in the way we actually haven't prevented it at Alithea.
3:18
It's ultra scalable in terms of costs.
3:20
So we're really in a way optimised the procedures, the reagents, some of the reagents we actually produce them in house as well to be able to offer extremely scalable costs.
3:31
And by scalable, I mean that the more it's done with DRUG-Seq, the lower the unit cost, the per sample cost actually gets.
3:39
And the ultimate advantage is that it's an unbiased approach.
3:44
So you can actually measure everything that is expressed on all the mRNA transcripts without actually having to predetermine what you're measuring.
3:55
Just to give you really a brief overview of DRUG-Seq in a nutshell.
3:59
So DRUG-Seq typically starts with frozen cell plates.
4:05
So it's extremely compatible with all existing cell biology and treatment part.
4:15
DRUG-Seq is compatible with 96, 384 and 1536 well plates.
4:19
And the first key aspect of DRUG-Seq is that it's an isolation free procedure.
4:26
So there's no need to isolate RNA through any type of magnetic bead or column purification.
4:32
The way the protocol starts is by basically just adding a lysis buffer that we offer on top of the frozen cells.
4:39
This basically solubilizes the RNA into solution, which then can be basically directly implemented in the subsequent steps.
4:48
And the second aspect, key aspect of DRUG-Seq is that the procedure is massively multiplexed, meaning that we can actually tag, at the very beginning, all the samples in a sample specific manner, after which we can basically pull all the samples in a single tube and process them in a tube for the rest of the workflow.
5:09
This drastically decreases the amount of reagents and time and tips that actually need to be used to run the workflow.
5:18
And this simplicity is actually what makes it unique in the sense that it's the only RNA-seq workflow in our knowledge that can actually be implemented at the scale of, for example, 1536 well plates.
5:33
It's very compatible with all kinds of next generation sequencing solutions.
5:38
This has been validated on Ultima Elements in Illumina and even MGI and the data analysis is, of course, open source, there's no nothing proprietary in the way we analyse the data or we provide the data.
5:51
Everything is open source, and we typically also provide support with the data analysis as well.
5:58
DRUG-Seq can be accessed in different formats for those that actually have capacity in house in terms of NGS and library prep, we offer kits with all the reagents needed to go from cells to sequencing ready libraries.
6:12
We offer end to end services as well in which you basically just need to receive the frozen plates.
6:17
And we take care of everything from library prep, sequencing and data analysis.
6:22
And recently we also started partnering with CROs to actually offer an end to end solution to really accommodate also the cell biology and treatment part.
6:35
Here I just wanted to show a couple of case studies of projects that we have run in the past that we were actually allowed to discuss, just to give you an idea of what's possible with DRUG-Seq and what's, what type of analysis and data can be generated from this.
6:49
This was one actually of the first DRUG-Seq projects that we have run in collaboration with the Broad Institute.
6:56
This was an interesting project because they were studying compound combinations.
7:00
So we actually profiled over 3800 compound combinations for a total of around 11,000 samples.
7:08
We detected around 12,000 genes per sample and this was actually a very fast project that actually took only seven days from the moment we received the plates to the moment the data was delivered.
7:20
Here's just to give you an idea also about what type of analysis typically is possible with DRUG-Seq data.
7:26
It's the data is of course very abundant.
7:29
So it starts to look more and more even if it's bulk RNA sequencing data, it starts to look more and more like single cell RNA-seq data.
7:36
So typically what we see as the first layer of analysis is actually clustering approaches like UMAP, PCA, t-SNE and marker gene identification.
7:49
And then typically what one can do is to actually isolate a subcluster or a few clusters of interest and then zoom in and understand what's happening actually there in that cluster.
8:00
Specifically for this project, the idea was to actually understand what compound combinations where we're doing, how they were interacting and what compound combinations were actually bringing back the gene expression state to the baseline.
8:14
So specifically for this project, we actually had to do thousands and thousands of pairwise analyses which is not typically done if you're not looking into compound combinations.
8:25
Another interesting project that we have done in collaboration with our partners at Axxam was another interesting one.
8:34
The number of samples was actually bigger in this case and it was interesting because here we actually did DRUG-Seq on several cell lines without any compatibility issue.
8:45
So DRUG-Seq is very compatible with different cell lines.
8:48
We do not need to optimise the procedure or the lysis step for every different cell line.
8:54
It's it typically works right off the bat when with different cell lines.
9:00
The idea here was to understand and get early views on markers for toxicity for the compounds that they were being processed.
9:10
And in this case, you know, we started looking at specifically clusters, how the compounds were clustering and whether there was a compound specific or a cell type specific clustering effect.
9:21
And you see how the data actually can be quite sparse when you start having a such a large data set.
9:29
Specifically here we started seeing a compound specific effect.
9:34
You see that actually the samples are actually clustering out with higher dosage of the compounds.
9:43
And again, one example of the analysis that we have done in this case was to actually check what compounds and what cell lines were actually reacting to different compounds.
9:52
So here in this case, for example, we started looking for specific signs of mechanism of action for the cell lines you see here, for example, this specific cluster, we could actually identify this compound that was actually setting out this cluster at any dosage.
10:12
So we didn't see, you know, any specific dose related response.
10:16
And we saw how this compound was actually acting differently based on the on different cell lines.
10:22
Whereas this compound was actually acting more gradually.
10:27
So it was actually giving a more dose response to the cell lines.
10:35
DRUG-Seq is one of those technologies that in a way require a conscious shift in the way that the investments are done in drug discovery.
10:46
It's an investment that is really done in with the view of increasing the amount of data that you can actually digest and play around with at the early stages of drug discovery.
10:57
And this is that this is the type of data that we see can really bring a benefit down the road because it allows you to take decisions faster and it can bring significantly advantages like the identifying liabilities early, accelerating go no go decisions, smarter target selection.
11:15
And also you can actually export and extend the knowledge that you have accumulated early in the discovery phase to actually clinical trials, which can bring despite the relatively small investment at the earlier phase, much faster decisions and effective decisions down the road.
11:38
Here I just wanted to mention also a new technology that we have just developed.
11:42
We're about to launch it actually in July.
11:44
And it's a new system.
11:47
It's I would say a big bat of Alithea’s and we're looking for partners at the moment that are interested in the extra layer of information that this technology will bring.
11:57
The technology is called Total DRUG-Seq.
11:59
And the idea in this case is that Total DRUG-Seq brings the same advantages as DRUG-Seq.
12:06
So it's extremely high throughput, it's scalable and low cost.
12:12
But in addition to three prime RNA sequencing data, so in addition of gene, in addition to gene expression data, we can actually profile long noncoding RNAs as well and full in a with full length level information.
12:24
So we can actually sequence everything across the RNA molecules here.
12:32
Just to show that Total DRUG-Seq is indeed a full length approach.
12:37
Typically Total DRUG-Seq was compared with standard non multiplexed and non-extraction free approaches like NEB.
12:45
And as you can see here, Total DRUG-Seq really in a way profiles entire length of the RNA molecule.
12:53
As I was mentioning, Total DRUG-Seq brings another additional layer of information.
12:57
So we actually in terms of genes detected, we are actually at par with DRUG-Seq.
13:02
We can actually detect the same number of genes, but when we start looking at specific transcripts, we can detect significantly more transcripts.
13:11
So we can actually start doing transcript level analysis and the partnerships actually we're looking for at the moment is to understand how important this additional layer of information can be in the discovery phase or any analysis that might be interesting to do with Total DRUG-Seq.
13:29
One example of an internal analysis that we started doing was to actually understand whether this data could actually provide insight into alternative promoter activation.
13:41
Here we considered a model in which HUH-7 cells were treated with different doses of TGF beta and we started looking at we basically profiled these cells with standard DRUG-Seq here in green and Total DRUG-Seq in different shades of blue.
13:56
You can see here that with standard DRUG-Seq, we can actually, the only thing that we can say here really is that the gene is actually expressed, which is perfectly fine when we're looking specifically at differential gene expression analysis.
14:08
But we, if we didn't do Total DRUG-Seq in this case, we basically would be missing out on a potential interesting biological phenomenon, which is this alternative promoter activation.
14:19
Here you see that basically promoter number one is activated when with lower or no TGF beta doses, whereas with a gradual increase in the dosage this was actually provided on this on 96 different samples.
14:36
You see that the second promoter gets activated.
14:39
And this is actually, we found it interesting, because promoter 2 is of course associated with is a hallmark of cancer and tumour progression.
14:49
And this was again, a type of biology that would not be identifiable if someone was actually looking only specifically at three prime RNA sequencing data.
14:59
So this is, you know, just to give an example of analysis that we started doing.
15:02
If you know if this sounds interesting, if this is the type of data that you would like to explore, please reach out because we're very interested in partnering and actually looking at the potential for this data together.
15:16
We did say thank you much for, thank you very much for your attention.
15:19
Our booth is basically just outside the room.
15:21
So if you want to know more, we have our team with Daniela Sofia to talk to you guys.
15:26
I'll be there as well.
15:27
And I think we also have 5 minutes for questions, if any.