0:00
Thanks so much, good morning everyone.
0:03
First of all, sorry because I have the task of waking you up this morning.
0:07
That's not the nicest one.
0:09
If I don't manage to make it, you can pass by our booth #33 and we can have a more personalised conversation about any of our products.
0:18
So my name is Fer, I'm the NGS manager in Europe.
0:23
And well, today I'm here to talk about Twist.
0:26
I think Twist doesn't need such a long introduction anymore because we've been in the field already for quite some time.
0:33
Just very briefly, in case you've never heard about us, our core technology is DNA synthesis.
0:39
But unlike everyone else, we don’t use 96 well plates.
0:42
We have created this silicon platform that allows us to really scale up the DNA synthesis, meaning that we can create this DNA much faster.
0:52
It is better quality and that helps you drive cost down.
0:56
So what is special about Twist is the way that we use this DNA within NGS.
1:02
We focus on target enrichment.
1:04
So as probably you know, sequencing is very expensive.
1:09
So what if you didn't have to sequence absolutely everything that is within your sample?
1:14
What if you could just choose those targets of interest and focus your sequencing reads there?
1:22
As an example, if you were interested only in the exomic content, this is approximately between 1 and 2% on the of the entire human genome.
1:32
If you wanted to do 20 reads over every one of your target and you perform an enrichment experiment with 7.5 gigabases of data, you would be meeting your goals.
1:43
If you're doing holding on sequencing, you would need 90 gigabases of data, so obviously you can see the drastic data reduction in order to achieve the same goals.
1:54
What this means for you as a researcher is that you can use your sequencing run to load many more samples to acquire much higher sequence in depth and this allows us to enter many more applications and meet more customers’ expectations.
2:12
So we initially started by specialising on the design of these enrichment probe panels and then with time we also optimise a library prep and target enrichment workflow that would allow you to get the best performance when using our probe panels.
2:30
So pretty much we can offer everything from the moment that you have your DNA until the moment that you have your enriched library.
2:37
And then you can use any sequencer that you wish, any analysis provider in order to extract your conclusions.
2:44
Basically that has enabled us to enable so many applications.
2:47
Pretty much anything where you can do sequencing, you can do also NGS target enrichment.
2:53
So today I don't want to go too deep into our projects, I mean into our products.
2:58
I thought that I would rather share with you some case studies so that you see how you can use your products because otherwise it gets very abstract, like what does this mean?
3:08
What is probe target enrichment?
3:11
So I'm going to be talking about three of them, one epigenomics, one transcriptomics, and one genomics case study.
3:17
Let's start with the epigenomics one.
3:19
Let me give you a little bit of background on this.
3:21
This is about imprinting disorders.
3:23
These are rare disorders caused by the malfunctioning of imprinted genes.
3:28
And those are genes that are expressed by pattern of origin, meaning that whatever you get when you're born from your dad or from your mom is the way that those genes are going to be expressed.
3:40
One of the mechanisms regulating this gene expression is DNA methylation.
3:46
So I guess you probably know that by methylating cytosines in CPG islands in the promoter regions or downstream the genes, you can finally regulate the gene expression.
3:57
And by alteration of those patterns of methylations, you can alter that going from a gene that is actively being transcribed to a gene that is inactive.
4:09
So we know of 48 locus today that are involved on that are expressing a pattern of origin manner and 17 of these are involved on disease.
4:22
They have been directly linked to disease.
4:24
So obviously it will be great if we could actually assess the methylation status of those genes.
4:29
What is the problem?
4:31
So far there were three main techniques used.
4:34
One of them is MLPA, which is very difficult to scale because it measures specific locus.
4:42
Another one is arrays which is not sensitive enough.
4:44
Another one is epigenome whole genome sequencing, which is great but super expensive.
4:49
So these researchers got to know about our methylation detection system, and they decided to give it a try.
4:57
What is this?
4:57
Basically this is a way of doing sequencing in a targeted approach.
5:01
We have developed an algorithm in order to design probe panels that will allow us to capture converted material after conversion.
5:13
What does this mean?
5:14
So basically you can start your prep from low input amounts, but you are also going to be able to conserve a load more complexity, meaning you're going to be reaching better results.
5:27
So we have partnered with NEB to provide the full workflow.
5:30
Basically you're going to be converting non methylated Cs to Ts.
5:35
Then you will do the capture with this approach that I told you where basically you have 4 probes per target, two against the fully methylated, two against the fully unmethylated, so you maximise the probability of pulling it down.
5:47
So this specific author generated a custom panel called ImprintCap against these 48 differentially methylated regions.
5:58
They were able to detect in 41 of those 48 regions with a sensitivity of 30% mosaic the methylation status.
6:09
And over here what you can see is part of the validation with six of the patients they tested, they found really high correlation with MLPA, which is the traditional method to test it in individual locuses.
6:22
And not only this, they could even expand it, for instance, to do CMB calling.
6:26
So obviously we just want to say you get a lot more granularity of what's going on in the genome of that patient.
6:34
So this takes me to the next case story, which is somehow related because it's also about gene expression, but this time is on transcriptomics.
6:42
So this is the BAF complex.
6:46
The BAF complex is also involved in regulation of gene expression.
6:52
The BAFopathies are the scissors related to anomalies in any of the components of the BAF complex that makes them not work properly.
7:01
And over here you can see one foetus which pregnancy was stopped at 21 weeks basically because it was considered not viable.
7:13
And it has a clear phenotype of serious coughing syndrome which is one of these bafopathies.
7:21
But on top of that, they found this hydrocephaly that they've never seen before related to bafopathies.
7:28
So they wanted to understand what was going on with this baby, why was it not developing properly.
7:33
So the first thing these authors did is actually do DNA exome sequencing to assess whether they could find any anomaly in their DNA.
7:42
And they did find these four base duplication in, ARID1A gene.
7:47
This is one of the components of the BAF complex.
7:52
So the next thing they thought after studying this duplication is that very likely this was generating an mRNA that was escaping the non-sense mediated mRNA decay.
8:04
This is a surveillance pathway that basically is going to degrade all the mRNAs coming from abnormal genes either because they have early stop codons because they can generate like aberrant proteins.
8:18
So what's going to happen is that it's going to be identified and it's going to be chopped.
8:21
So you cannot create the protein.
8:24
So they thought how to assess this.
8:26
Well, let's check if we can actually find the mRNA in these cells, in the cells of this baby.
8:34
So for that they did RNAseq with enrichment just to focus the reads on those exons of interests.
8:41
And they did find that this duplication was also seeing the mRNA, highly supporting the hypothesis that it is indeed escaping this NMA surveillance pathway.
8:53
How did they do this?
8:55
Well, they use RNA seq solutions.
8:57
These are very similar to our DNA solutions.
8:59
These are very fast workflows that pretty much just include the cDNA conversion ahead of the workflow.
9:07
The good thing is that we use random exomers for the cDNA generation.
9:12
That is going to guarantee you like good coverage of the entire body of the isoform.
9:17
It doesn't have the bias that you would usually get with regular Poly A generation.
9:26
On top of these you can do enrichment after you have the library prep generated which will allow you to refocus your reads on those targets of interest.
9:34
You can see these over here how with much lower amount of reads you can get similar coverage of your targets.
9:45
So for our RNA solutions we have the sign and novel approach to probe design which is called the Exon Aware algorithm.
9:53
And basically this consists of avoiding these probes that are spanning exon boundaries.
10:01
And what you get when doing that is really enhancing the representation of any isoform, any fusion that you might have.
10:09
Otherwise, you will have biases for more abundant isoforms and more abundant fusions.
10:18
And obviously apart from that, you're going to be really enhancing the representation of in this case with the RNA exon, your exonic content of mRNAs and just getting rid of all the noise and the background that other type of RNAs might be introducing in your experiment.
10:37
So finally, I'm going to be talking about my third case study.
10:41
This is about liquid biopsy pretty much assessing within bodily fluids what is the content of DNA coming from cells, whether they are healthy cells or in this case cancer cells.
10:59
So these cells are going to be throwing these pieces of DNA to your cfDNA and you can actually monitor these DNA within your bloodstream in this case.
11:11
So for instance, for MRD minimal residual disease, the aim would be once a patient has cancer, you obviously are going to have increased levels of these markers like cancer region markers in the bloodstream.
11:26
If there is treatment, these markers will go to minimum levels, but then you can monitor and assess whether the disease is reoccurring or is maintained stable.
11:39
And what's the benefit of using this liquid biopsy?
11:42
It is a non-invasive test.
11:45
You don't need to take the patient into surgery to get a sample.
11:48
So obviously the benefits for a patient are huge.
11:52
So this company from France [unclear], they partnered with us.
11:56
They wanted to test whether our cfDNA workflow could allow them to have MRD monitoring.
12:03
So first of all, what they did is assess the sensitivity that they could get when using our workflow.
12:08
Basically depending on the input mass, just so that you know, whenever you're working with cfDNA, the masses are usually really low and that can complicate the sensitivities.
12:20
But they found that even when you see one nanogram of DNA, they were able to at least detect 2 mutations from the disease signature and it was allowing them to have a 0.003 percentage variant infrequency.
12:36
Obviously they had to do some studies to assess the limit of blank.
12:40
So what is the probability of having false positive?
12:47
And they did this with our cfDNA standards, which are just like standards with variants at different levels of variant infrequency.
12:56
They were also interested to understand what are we seeing from the real sample.
13:01
Why is this so whenever you are processing a sample for NGS, you're going to be losing material.
13:08
The main step where you're losing material is doing the ligation step because ligases at the end of the day are enzymes that do not have 100% effectivity.
13:17
So for these workflows, we have engineered a ligase to maximise the library conversion, meaning getting the majority of the molecules within your input sample converted to libraries that can be sequenced.
13:29
And what they found is over 75% library conversion rate in all cases, which gave them like peace of mind because obviously what they are getting at the end of the results is really representative of what they are having in their input sample.
13:46
Finally, I want to show you the way they tested these workflow with real patients.
13:53
So B1 basically is a sample from the patient before treatment.
13:58
B2 is a sample from the patient after treatment, like immediately after treatment and B3 for weeks after treatment.
14:05
And they assess different markers from the cancer cells, and they saw how whenever they went and proceeded with the treatment, the levels of those markers really decreased.
14:19
And sometime afterwards, in some instances, those levels started to increase again.
14:24
So really reflecting what we know from the biology of these patients that it was happening.
14:29
Same with the fraction of ctDNA, ctDNA will be the tumoral fraction within the cfDNA.
14:36
They saw that following the same profile.
14:39
Really like supporting the use of this workflow for MRD assessment.
14:45
And this is all that I have to share with you.
14:47
I hope I didn't fly too much.
14:49
I hope you could understand that.
14:50
If you have any questions, feel free to bring it up now or otherwise just pass by our booth.