0:00
It still has its own challenges.
0:02
There are non-coding transcriptome or the RNA mutations that have no impact on the protein, therefore don't have any direct impact on the phenotype.
0:13
For the longest time everybody knew that the proteome is the magic key for knowing the and understanding and predicting the phenotype, but it had its challenges.
0:24
There wasn't many high multiplexing available.
0:27
ELISA was available for the longest time.
0:30
Mass spec had his own limitation of the sensitivity.
0:34
So there weren't any high throughput, high multiplex technology that was allowing the customers and allowing the researchers to make a direct connection of what is happening in the biology and what is connected to the phenotype.
0:51
And here is where SomaLogic really started to step in.
0:57
Just a little background, we have been actually focusing on the human proteome for over 2 decades to be exact.
1:04
For 24 years, we have been imagining and envisioning the multiplexing of the proteomics in a different way compared to the traditional Mass spec or antibody because of its limitation.
1:26
If you look at the evolution of our technology and how we have come along, you can see that how we have gone from the thousands of the protein all the way to what recently?
1:48
How in last December we launched our newest panel 11,000 proteins that cover over half of human proteome.
1:59
What are the advantages and the result of this menu expansion?
2:08
Why do we keep focusing on expanding that and casting a wide net.
2:13
We make sure that we are allowing the researchers that are after biomarker discovery and doing the multiomics, they have the highest power to layer the human proteome to the existing genomics and the transcriptomics that are out there.
2:29
We made sure that the composition and the distribution of the coverage of the protein that we have is mirroring what is in human proteome.
2:38
So we are not biasing one group of the protein as you can see here.
2:50
You can see how the distribution of the protein in human proteome is and how we make sure that we are almost mirroring that so that whatever coverage that you are getting is in parallel to what is available in the human proteome.
3:04
And in result of that, you can see we have over 50% coverage of all KEGG major pathways and how we have significantly increased the coverage of various diseases and the biological pathways.
3:20
You can see from our own 7 KEGG panel that was our previous version to the 11 KEGG that is right now how by just increasing this 4000 how the disease coverage for us significantly improved.
3:35
This is really showing why our mission is to expand our panel, why we keep going to make sure that we have a more and more coverage of the human disease because we truly believe you don't know what you don't know and what you are missing.
3:51
But how we are doing that, if it was easy why is not everybody is doing that.
3:55
And our secret sauce is what we call SOMAmer.
3:59
SOMAmer stands for slow off-rate modified aptamers.
4:03
In the little video that you can see, the grey matter is protein.
4:08
That teal colour is backbone of our SOMAmer, which is a single stranded DNA.
4:13
And we have a modification and this is what is proprietary to us.
4:17
We have a modification attached to this backbone of the single stranded DNA that goes on locking through 3D structure of the protein epitope.
4:27
Because we are synthesising this complex ourselves, we know that how we designed it in such a way that it doesn't easily come off and it go very specifically to the protein epitope.
4:41
And that's why we can go without compromising specificity, without compromising sensitivity.
4:47
We can go to such a high Plex of the protein.
4:51
What are the main pillars of our technology?
4:54
I have summarised here.
4:56
I am not going to go into the details of each of them but if there is any question, I can expand why and how we are doing that.
5:05
As I mentioned, we are covering over half of the human proteome.
5:10
We are the highest coverage of the human protein that is out there and we are the highest validated technology.
5:20
We don't dismiss other technologies.
5:23
Over 75% of our menu has been confirmed by mass spec.
5:28
We know that mass spec has sensitivity issues.
5:31
So for example, if you have a high abundance protein and low abundance, it gets saturated.
5:37
But we have found a way around that to confirm our protein target using mass spec, we confirm with the CSPTTL, we confirm with the antibody whenever that is available.
5:50
And that is why we are saying that we have a higher specificity of not just a primary validation, also orthogonal validation of our technology, something that is quite unique to us is correlated data.
6:04
And I'm going to explain what does that mean.
6:08
Many of you have heard about absolute quant, how absolute quant is essential for clinical trial.
6:15
And the reason for that is that you are comparing your signal to the standard care and you are anchoring it to a reference.
6:24
However, if you have ELISA two different from two different company and their standard curve are different and their buffer are different, your absolute quant is not the absolute quant, it's a form of relative quant.
6:36
So we got around that, how we did that we have actually generated over thousands of samples from the diverse background and the data that we have generated are going to get used as a reference population when we are normalising your data.
6:51
Therefore, if you run a sample today, six months later, or a year later, or four years later, you can combine the data and compare the data directly with each other.
7:01
So we reimagine what the absolute quantification actually means.
7:08
We have the ability to scale up.
7:10
This is very important for the people that are using ELISA or Mass Spec.
7:14
You can run the 8 plates of 96 format in a week very easily.
7:22
So you can scale up very fast.
7:25
We require very low sample volume and the diversity of the samples from the various animals to the various sample matrices.
7:34
You can use our technology very large dynamic range from femtomolar up to the micromolar in the same sample without any additional treatment.
7:44
You can detect all of the proteins and something that we are very proud of and if you go to our website, you see that we keep talking about how reproducible our signal is.
7:56
Our CV between the batch inter plate CV is 5%.
8:03
Because of all these unique features that we have.
8:07
We didn't stop by just offering the discovery to you, not just offering that you cast the wIDE net and look at all the proteins, finding which protein is associated with your disease.
8:19
Of course you can do that.
8:21
That's the very first step.
8:22
And you can choose from the target panels that we have or you can run the 11,000 protein that you have or you can use the custom.
8:30
You can go in our menu, pick and choose whatever protein and create your own 1500 custom panel.
8:36
But we didn't stop here because how our data is cleaned and I talked about how our data is global and doesn't need any bridging normalisation and how our CV is good.
8:48
We took it to the next step which is what we call a SomaSignal test.
9:00
This is where we truly took our signature to the clinical space.
9:06
We started by using the AI power and machine learning using collaboration with the various large pharmaceutical companies that they provided this clinical sample for us with the clinical information.
9:19
We use that to associate the protein signature with various diseases and we'll be able to predict those diseases years in advance.
9:31
You can see here that in our menu there are two examples of the offering that currently we have.
9:39
Of course many of our customer have designed their own AI and SomaSignal test and they are offering it as a clinical test.
9:46
But this is what we have done and we have gone through the regulatory test.
9:50
One form of the SomaSignal test is sample handling assessment.
9:55
This is something that a lot of people that are working with protein, they actually find it as a pain point.
10:01
If you look at the protein samples or plasma wrong, something happens in the protein.
10:07
If you leave it in the bench for two hours extra, some of the protein starts to degrade or start to show a different affinity to binding.
10:16
The sensitivity of the sample handling for protein for decades has been a challenging part that makes the data analysis hard.
10:26
So we found a way that we associate certain protein signature with the various sample handling and we provide that information to you.
10:35
So once you run your sample, let's say that you have samples that have been centrifuged in various times, some of them one hour, some of them 24 hours.
10:42
We tell you that how they have been handled so that you can assign a confounding variance in your analysis without eliminating those samples.
10:52
You can tease out those signatures out, you can tease out those contaminants out of your actual analysis.
11:01
So that is one of the SomaSignal tests.
11:04
The other one that we have is the clinical risk assessment and you can see the example of that.
11:09
The menu constantly keeps growing that for various diseases we are able to give risk scores or the prediction of the outcome of these diseases.
11:20
For example, for the dementia test, that is our newest test, up to 20 years in advance, we can tell you if anybody is going to develop dementia or not and if they are at the high risk score or not.
11:34
And if there is a life intervention, how the risk score is going to change.
11:39
We have published this data and if anybody is interested, I'm more than happy to send the publication.
11:43
Many of these risk scores are going to the regulatory approval right now too, but I want to take one of these ones to show you how actually these tests has been developed and how accurate they are.
11:58
I'm going to take the Nash test and I am going to jump into the paper that just was published a couple of months ago in Lancet.
12:08
This is a study that was commissioned by Litmus Consortium.
12:12
There are various groups that collaborated together, shared the samples and these are the samples that by the liver biopsy they were confirmed that they have the Nash disease or the advanced fibrosis.
12:29
So they had the group of population that already have been confirmed and the group that was controlled.
12:35
They have tested 17 different clinical tests including the biomarker and imaging methods that are currently used for the Nash disease, and they have put them by the calculating sensitivity and the specificity.
12:53
They gave them the score of the area under the curve AUC to say that which method has the highest accuracy for detecting the protein.
13:03
And you can see that out of all the 17 methods for both of the categories, SomaSignal test outperformed all of those existing methods.
13:14
Also you can see that in the graph that is on the side here.
13:20
OK, in the graph on the side, you can see how many samples are needed to have the confidence of the 95% for the disease prediction that you have.
13:34
And this is because the liver biopsy 33% of the time it fails the prediction.
13:39
So it has a high failure.
13:41
So you have to have actually pretty high population to make sure that you have a good prediction.
13:46
And you see that out of all of them, this is the lowest sample number that you need.
13:52
You need actually only four to get the 95%.
13:56
So not only does it have the highest ranked in the accuracy, it has the highest diagnostic ability as well.
14:05
This is a paper that again, if you're interested, I am more than happy to send it to you.
14:10
This is something that outside of SomaLogic has been done by the consortium to make sure that the how accurate the current method that is out there is.
14:25
I just want to remind you that my presentation is short, but we have all offers and all the features of our technology that can cover from the research all the way to the diagnostics.
14:40
So if your research is fitting in any of that, there are the features in our technology that can assist you.
14:46
Before I wrap up, I just want to let everybody know that we have actually the grant award competition that is going on right now.
14:56
It is open to everybody in the UK.
14:59
If you would like to win 40 samples free of the charge or the 2 runners up are going to get 50 and 25% discount.
15:08
You can submit your grant right there.
15:10
You can scan also the code to take you to the website.
15:15
And thank you for listening.