Diagnostic odyssey, cases solved by whole exome sequencing

0:00
Good morning, good afternoon and good evening and welcome to today's educational webinar titled diagnostic or you see causes sold, cases sold by whole exome sequencing.

0:16
The webinar is brought to you by Blueprint Genetic, a genetic knowledge company committed to providing an innovative approach to genetic testing and to ensure accurate and confidence in your clinic clinical practice.

0:33
My name is Tina Boppio and I have the pleasure to host you today.

0:38
Please submit any questions you may have in the questions book box.

0:44
You can submit them throughout the webinar and we will answer as many as possible at the close of the webinar.

0:52
We are excited to have Doctor Kirstie Wells as our presenter today.

0:58
Doctor Wells is a Senior Genetist at Blueprint Genetics specializing in interpretation of ophthalmology panel and whole exome signals in data.

1:10
She has a background in both research and diagnostics.

1:15
Before joining Blueprints in 2018, Kirsty completed PhD and postdoctoral research fellowship and undertook in depth training in genetic diagnostics in the UKUKS National Health Service.

1:32
Kirsty is AUK certificated clinical scientists.

1:38
So thank you for being here today, Kirsty, please.

1:43
Thank you, Tina.

1:44
Thanks for the nice introduction.

1:46
And hi, everyone.

1:48
Good afternoon.

1:50
Good, good morning.

1:51
Depending on where in the world you are, Thanks so much for coming to spend an hour with us today to talk about whole exome sequencing cases.

2:00
As Tina said, my name is Kirstie Wells.

2:02
I'm a senior geneticist at Blueprint Genetics and a whole exome sequencing team lead.

2:08
So it's a pleasure today to get to talk about my favorite topic, which is obviously whole exome sequencing.

2:18
So our field of rare disease diagnostics is, is, is really moving towards whole genome sequencing, but whole exome sequencing still remains really relevant when it's high quality and it's optimized to achieve the best possible diagnostic yield.

2:36
So that's what I want to try and convince you of today, the continued relevance of whole exome sequencing in a world in love with with whole genomes.

2:44
And this is especially for patients on a diagnostic odyssey.

2:49
So for this webinar today I've I've got some interesting patient cases to show you woven in with a bit of technical information.

2:57
And I hope you find this this interesting and educational.

3:03
So first of all, I'll begin with a quick introduction to Blueprint Genetics.

3:08
At Blueprint, we deliver high quality genetic testing to the global clinical community.

3:14
We have over 250 employees these these days and we're serving over 4000 clinicians in 70 countries.

3:23
And we have core operations in Helsinki in Finland, where I'm talking to you from today.

3:30
And we offer a range of different NGS based test types depending on the customer's needs.

3:36
So we offer gene panels.

3:38
Of course, we also offer a whole exome sequencing, single gene testing, variant specific testing and also screening tests.

3:49
And in this webinar today I will focus on whole exome sequencing.

3:57
So let's start with some background to whole exome sequencing to get us as all on the same page.

4:04
Whole exome sequencing or or Wes for short, basically means aiming to sequence the protein coding regions of all known genes.

4:13
That's about about 20,000 of them.

4:16
So in a Wes around 1 to 2% of our of our total DNA is sequenced because of course the majority of our DNA is non coding because there's no restriction on what genes are sequencing a wears.

4:33
And unlike in a panel, this means that wears can be a very powerful diagnostic tool indeed, and it's especially useful for patients with complex phenotypes or an unclear clinical diagnosis.

4:46
So a patient like this may have had previously inconclusive genetic testing because it may have been difficult to choose what genes to sequence for that patient.

4:59
So that means that a patient like this can end up on a diagnostic odyssey with sometimes many years of negative and inconclusive investigations.

5:12
So where's is a very powerful test, but there may also be limitations of where's as a test in general, depending on how the assay is designed.

5:24
So where's as a general definition is sequencing of the exome.

5:29
So that means first of all that non coding regions are typically not covered.

5:37
The mitochondrial genome may not be included in the assay somewheres assays in some labs do cover MiTo and some don't.

5:46
Also important to be aware that difficult to sequence regions across the exome can be challenging and this is of course the case for other other NGS based test types as well.

6:00
CMV detection can be variable between different assays from different labs and and particularly small CMVS can be challenging to detect.

6:10
And then finally a a challenge with Wes is how to deal with with all those variants and how to get to the relevant ones efficient efficiently.

6:19
So typically after whole exome sequencing a patient has about 35,000 variants to deal with.

6:26
So it's important to consider how to do the analysis and the interpretation to get efficiently from that huge number to the truly relevant reportable variance for the patient.

6:41
And it's really important to be aware that that all Wes, Wes assays are not necessarily the same in terms of these limitations.

6:50
And this is really important because how how the potential limitations of Wes are addressed by a particular assay design can really influence whether a patient receives a molecular diagnosis from the testing or not.

7:08
So as a clinical lab geneticist, I like nothing better than interesting patient cases.

7:14
So cases where we found the answer for the patient and being able to help the family.

7:20
So in today's webinar, I will show you 6th interesting worst cases.

7:26
These are based on on real cases from the Blueprint Genetics Archive, but with details changed and removed to protect protect patient confidentiality.

7:36
And each of these cases aims to show how it's possible to try to tackle some of those potential limitations and challenging challenges of Wes that I that I was talking about before.

7:51
And that by trying to tackle these limitations, this can really make make the difference in terms of finding the answer for the patients and especially for those who have been on a diagnostic odyssey.

8:04
OK, so let's get started with the cases.

8:06
And in the first case, we are checking the introns.

8:12
So this patient was an adult with a clinical diagnosis of neurofibromatosis type 1.

8:20
And we were told that previous testing elsewhere of the gene associated with this disease, NF one, had been negative.

8:31
So for this case, let's take a moment to think about where in a gene we might find disease causing variants.

8:37
So here we have a really simplified gene model showing exons, introns, splice regions and UTR's.

8:46
So we most commonly find disease causing variants within the exons, which are the protein coding regions of a gene.

8:57
We also frequently find disease causing variants within the splice regions.

9:01
So these these are the nucleotides either side of an Exxon intron junction and variants here can often disrupt splicing.

9:13
So these two regions, the exons and the splice regions, these are the ones that are typically covered by whole exome sequencing.

9:22
However, it's also possible to get disease causing variants in non coding regions such as deep within introns and also even within UTRS or the promoter region.

9:39
So in our patient with with NF1, what did we find?

9:44
We detected a heterozygous deep intronic variant in the NF 1 gene.

9:49
This is a pathogenic variant that has been reported in the literature in multiple individuals with NF1 related phenotype and it's been shown in splicing studies to lead to cryptic exon insertion.

10:04
So if we cast our minds back to the last slide, we learnt that an exome would cover typically the exons of a gene and and the splice regions and in terms of the splice regions going 20 base pairs into the intron from the Exxon intron boundary to pick up the splice region variants.

10:25
So a standard whole exome sequencing assay just covering the exons and the splice regions would not detect this, this deep intronic variant, it's too far into the intron.

10:37
This region would not be covered.

10:40
But the blueprint was has is boosted with oligos to enable detection of variants in in in custom targeted non coding regions where there are clinically relevant variants.

10:57
And so this deep intronic region in NF one is one of those custom targeted regions in our assay because we want to we want to be able to detect this known pathogenic variant because because it's known to be clinically relevant.

11:15
So this way we can detect almost 2000 clinically relevant non coding variants on our on our wares, including this NF 1 variant.

11:29
So summarizing this case, this patient received a molecular diagnosis of neurofibromatosis type 1 ending their diagnostic odyssey.

11:39
Because if you remember, previous NF1 sequencing in this patient had been negative.

11:44
And of course we don't know the nature of that testing, but I would guess that it did not include sequencing of the non coding regions.

11:53
So that's why this pathogenic variant was was not detected previously despite the fact that the NF 1 gene had already been tested.

12:01
So this result means that no further investigations are required for this patient and it's very useful for genetic counselling and management.

12:12
And this is thanks to the inclusion of clinically relevant non coding variants on the Wes assay.

12:20
So this is the first way in which we try to meet one of the major limitations and challenges of Wes by boosting the assay to detect clinically relevant non coding variants.

12:35
OK.

12:35
So let's move on to the next case now, which looks beyond the nucleus.

12:43
And this case involved a child with seizures, developmental delay and cardiomyopathy.

12:53
And for this case we're going to talk about the mitochondrial genome.

12:57
So let's start with a with some background to that.

13:02
So the mitochondrial DNA is a circular molecule which is just over 16.5 kilobases in size and it contains just 37 genes which are important in cell function.

13:17
It's important to note that mitochondrial DNA is maternally inherited and also that the mitochondrial genome is is polyploid, meaning that there are several 1000 copies of the the mitochondrial DNA in each cell.

13:33
So this means that you can have mixed populations of mutated and wild type mitochondrial DNA within a cell or even within a tissue or individual.

13:45
And this is an important concept called heteroplasmy and pathogenic variants in in in mitochondrial DNA genes can cause diseases with a very wide range of clinical Spectra, very variable presentations and variable ages of onset.

14:07
So this makes them challenging to diagnose.

14:14
Historically in diagnostic labs, the molecular diagnosis of of disease caused by MiTo DNA variances is typically involved screening for specific variants using PCR based methods or Southern blotting.

14:29
But these days NGS has become the gold standard for molecular diagnosis of mitochondrial disease.

14:37
Because with NGS you can look at the whole mitochondrial genome rather than just specific variants.

14:44
And also you can include MiTo DNACMV analysis in parallel with the sequence sequence analysis.

14:53
And also you can detect potentially low lower level heteroplasmy.

14:59
And all of this can be performed together with nuclear analysis to maximize diagnostic yield.

15:09
So if I if I can take you back briefly to this introductory slide where we went through how Wes is a powerful diagnostic tool, especially for patients with complex phenotypes, unclear clinical diagnosis and inconclusive previous genetic testing.

15:26
All of this perfectly applies to patients with mitochondrial disease.

15:30
So these phenotypes can be very complex, very variable.

15:34
The diseases can be hard to diagnose.

15:36
So that means actually that that Wes it is a really perfect test for this kind of patient.

15:42
And so it makes perfect sense to include mitochondrial DNA analysis on on a Wes test.

15:49
So that's why MiTo DNA sequence and CMV analysis is now included on the BPG Wes.

15:59
So going back to the patient in our, in our case the the child with seizures, developmental delay and cardiomyopathy, the mitochondrial DNA analysis on the Wes identified an interesting missense variant in the mitochondrial DNA gene MTATP 6.

16:20
Also, we see that the patient was homoplasmic for the missense variant identified and we also had the parental data in this case and the variant was not detected in the mother's sample.

16:33
So this was consistent with the variant possibly being de Novo in the in the index patient.

16:42
So this gene MTAT P6 NK encodes ATP synthase 6, also known as complex 5.

16:51
And this this is a subunit of ATP synthase, which is the enzyme that is instrumental in producing cellular energy And pathogenic variants in this gene are associated with with highly variable associated phenotypes as is very typical in mitochondrial DNA diseases.

17:15
So the phenotypes frequently reported in patients include things like developmental delay, seizures and retinopathy.

17:25
But there are many other less typical presentations as well, as you can see on the slide.

17:32
So phenotype matching for mitochondrial diseases can be a bit challenging because of this phenotype variability and breadth.

17:42
It can be hard to match what your patient has with what's described for, for the disease associated with, with, with, with the gene because because the, the, the phenotypes are often so variable and broad.

17:55
But at least in this case, our patient is reported to have developmental delay, seizures and cardiomyopathy.

18:05
So this gene is certainly relevant for our patient.

18:11
So the next question is what is the classification of this variant that we've identified and and how do we actually classify mitochondrial DNA variants.

18:22
So we do have a a specifically designed classification scheme for mitochondrial DNA variants, which is is based on on ACMG guidelines.

18:34
And in that scheme we consider first of all the frequency of the variant.

18:41
So the Nomad version three and Helix databases have mitochondrial variant frequencies and we utilize those and our MTAT P6 variant of interest is rare.

18:54
So it's actually absent in Nomad version 3.

18:56
So that's consistent with a rolling disease.

19:00
We also consider the phenotype match and as shown in the previous slide, we do have some phenotype overlap.

19:08
In this case, we consider whether the level of the variant is relevant.

19:15
So is it homoplasmic, heteroplasmic and what level of heteroplasmic do we find?

19:21
And this variant was detected as homoplasmic, so potentially very relevant in our patient.

19:29
We also consider reports of the variant in in patients in the literature.

19:33
And this variant had actually been reported in in multiple patients with MTAT P6 related phenotypes with high levels of heteroplasmy.

19:44
And in a number of cases, the variant was reported in the literature as apparently de Novo.

19:51
And including in our current case, our data is consistent with the variant being de Novo.

19:57
So we consider that good evidence for pathogenicity as well.

20:04
And then finally, we also consider the results of in silico tools and conservation of the affected.

20:12
Amino acid.

20:13
So this variant was predicted to be deleterious by the Insilico tool that we use for protein coding mitochondrial DNA variants and it also effects a highly conserved amino acid.

20:28
So bringing all that evidence together and applying our, our classification scheme for mitochondrial DNA variants, this variant was classified as pathogenic.

20:41
So this patient received a molecular diagnosis of, of mitochondrial disease.

20:48
And as the variant was apparently de Novo, this means that the recurrence risk is, is expected to be low.

20:55
Although we can't exclude the possibility that this patient's mother does have the variant at lower levels than we're able to detect, especially because we are only testing 1 tissue type.

21:08
We're only testing blood.

21:10
So we need to be a bit cautious with the interpretation there.

21:15
But nevertheless, this result is useful for genetic counselling and may inform future surveillance for this patient for other possible manifestations of mitochondrial disease.

21:30
And so the answer for this patient was found thanks to the inclusion of mitochondrial DNA on the West and also thanks to careful interpretation and and application of a of a systematic evidence driven classification scheme specifically for mitochondrial DNA variants.

21:55
OK.

21:55
So let's move on to case number three now, which we've called cover your bases.

22:03
And this case involved a child with poor coordination and ataxia and the with analysis identified a compound heterozygous variant in a gene called MSTO 1A frameshift variant and missense variant.

22:24
So we had this was a Wes family case.

22:27
So, so the, the, the patient's parents were included in the analysis and we could see that one of the variants was identified from one parent and the other variant was identified from the effort was detected in the other.

22:40
So this these two variants were in compound heterozygous state.

22:44
And based on our classification scheme, both of these variants are classified as likely pathogenic.

22:51
And this gene MSTO 1 is associated with an autosomal recessive disorder called MiTo mitochondrial myopathy and ataxia, which has a very good match with with the patient's phenotype.

23:06
But what's interesting about this gene MSTO 1 is that testing of this gene by short read NGS is known to have limitations due to severe sequence homology with a pseudo gene.

23:21
Let's have a closer look at this gene msto 1 then.

23:25
So MSTO one has a highly homologous non functional pseudo gene msto 2P and the overall nucleotide identity with the MSTO 1 functional gene is 99.5%.

23:43
And both of our variants of interest were found in Exxon.

23:46
Four of the gene where there is actually 100% homology with the pseudo gene.

23:54
And this kind of sequence homology presents a real challenge for NGS based sequence analysis for a number of reasons.

24:03
So in a typical bioinformatic analysis, sequence reads that align to several genomic positions due to homology tend to be discarded.

24:15
So this can cause gaps if if those reads have been discarded or low sequence coverage in those regions.

24:25
So, so that means if there are disease causing variants in these regions, then they might be missed due to due to no or low coverage in that region.

24:38
Or if there is a variant called in a region like this, it can be difficult to tell a variant that's come from the real gene from a from a variant that's come from the pseudo gene.

24:49
So if sequence reads containing a variant from the pseudo gene are mismatched to the functional gene, then that may result in a false positive variant call.

25:01
And then conversely, if sequence reads containing a functional gene derived variant are mismapped to the pseudo gene, then it may result in a false negative variant call.

25:17
So there are plenty of clinically relevant genes that are affected by sequence homology of different types and are therefore difficult to sequence by NGS.

25:26
So, so msto 1 is only one of many and some of those genes are shown on this slide, but there are there are many others.

25:36
So, so this is a really important limitation to be aware of in NGS based genetic testing.

25:42
And it's it's worth knowing if you have a particular gene of interest, whether that gene is affected by by homology.

25:54
But this limitation can potentially be tackled by careful assay design.

26:00
And the the most important components of this are longer paired end read lengths, which improves mapping quality, a custom capture oligo design for difficult to sequence regions, so enabling enabling better coverage in those regions.

26:22
Also also an optimized insert size to Fillmore low coverage gaps and also customized bioinformatics.

26:32
So a couple of years ago, we launched some improvements to our Wes assay where we applied applied these components and added new oligos and optimized the insert size.

26:46
And this has brought us better coverage, fewer gaps and better uniformity across across the Wes.

26:54
So this means that our the coverage of our wares is now on average 99.6% at 20X with improved coverage in difficult sequence regions and the depth of coverage is on average 154 X.

27:11
So I've just thrown some numbers at you.

27:13
So let's let's look a bit more closely at this topic of coverage and what these numbers actually mean.

27:19
So look, if you had one shot or one opportunity to seize everything you ever wanted in one moment, would you capture it or just just let it slip?

27:30
So that's a quote from the well known authority on Human Genetics, Eminem, and that's our target region.

27:38
So mean depth of coverage refers to how many times each nucleotide in the target region is sequenced on average.

27:48
And the number of times this sentence is repeated is our depth.

27:53
And then breadth of coverage refers to how much of the target has actually been sequenced.

27:59
So how what you wanted to sequence, you did actually sequence adequately.

28:04
So how much of this song lyric have we have we captured?

28:08
And typically we quote this as a percentage of nucleotides sequenced to a depth of at least 20 reads or or 20X we say.

28:19
And what you're looking for in the optimal with assay is for good even breadth of coverage together with good depth of coverage.

28:31
And it's important to realize that not all tests are the same with respect to this.

28:36
So we, we did an analysis of exome coverages from different providers and we see that just because the tests are named the same, so whole exome sequencing, it doesn't mean that the performance is the same.

28:50
So here on the charts, coverage ranges from 97% at 20X up to 99.6% on average for our current Where's assay.

29:02
And then it was 99.4% at 20X for the previous version of our where's assay.

29:08
And so bearing in mind that for Where's this is never 100%, this is that's a known limitation of where's, but you might think that these numbers actually look quite close to each other.

29:21
So 99% versus 97%, does this really make a difference?

29:29
And in fact small percentages like this actually represent very big differences when you extrapolate that out to the whole exome.

29:39
For example, in the in the Blueprint genetics exome, we have mean breadth of coverage of 99.6% at 20X.

29:47
And what that means is that approximately 120 thousand base pairs will be sequenced fewer than 2020 X so suboptimally.

29:58
And this equates to about 480 exons or or or maybe 53 genes.

30:04
So even though we try to get as close to 100% coverage as possible, this is a limitation of exomes that coverage is never 100%.

30:13
And, and this is what this limitation actually means in practice.

30:18
So then if we compare that to another lab that offers 97% coverage at 20X, what we find is that the number of base pairs that aren't meeting our desired coverage goes up to 900,000, which represents many more exons and, and many more genes potentially.

30:40
So that's actually a huge, huge difference even though these percentage differences may seem small.

30:49
So let's go back to our case with the two variants in MSTO 1.

30:54
And thanks to measures taken to improve detection of variants in difficult to sequence regions like this one, the depth and breadth of coverage of the MSTO 1 gene were good on the West assay.

31:07
And so we were able to detect these diagnostic variants by NGS and the variants were confirmed using a custom PCR SANGA confirmation method.

31:22
So we were able to confirm a molecular diagnosis of autosomal recessive mitochondrial myopathy and ataxia in this patient, which provided important information for genetic counselling in terms of risk to any siblings.

31:37
And also, patients with this disease as described in the literature tended to have slow or no progression of their symptoms and minimal cognitive involvement.

31:47
So, so this result may provide some useful and perhaps reassuring prognostic information for the family as well.

31:57
And this is thanks to optimization of coverage in this region that is known to be difficult to sequence by NGS, which allowed us to detect 2 diagnostic variants in this gene that that that may have otherwise gone undetected by other assays with with less coverage of this difficult region.

32:21
OK, So let's move on to the next case now, which reminds you to mind the mind the small gaps.

32:28
And this case involved an adult with developmental impairment and epilepsy and research where's elsewhere had been negative.

32:41
So on the blueprint genetics, whereas we identified a homozygous deletion of a single exon in a gene P gap one and the deletion was quite small, approximately 300 base pairs in size.

32:58
And the exon involved is out of frame, which means that this deletion is predicted to lead to a loss of function and loss of function is a known disease mechanism in this in this gene.

33:12
So that means that this deletion was classified as as likely pathogenic.

33:16
So this this patient is homozygous for a likely pathogenic deletion and both parents were heterozygous for the deletion because this was a a Wes trio.

33:30
So bialelic pathogenic variants in in in this GMP gap 1 cause a severe neurodevelopmental disorder and autosomal recessive disorder.

33:42
And the gene encodes a an enzyme which catalyses a step in GPI biosynthesis.

33:54
GPI is a phosphoglyceride that that plays key roles in a in a variety of of of biological processes and the disease associated with this gene is a neurodevelopmental disorder with dysmorphic features, spasticity and brain abnormalities.

34:12
And as said, this is an autosomal recessive disorder and some of the associated phenotypes include severely delayed psychomotor development and anesthesias.

34:25
So this was a match with our patient's phenotype.

34:31
So in this case, we detected a small CMV.

34:35
So let's let's now take a look at CMVS in general.

34:39
So a CMV is a region of DNA larger than about 50 base pairs, which shows a change in copy number compared with a reference genome.

34:48
And we know that CMVS are a significant cause of genetic disorders.

34:52
A study published in back in 2018 investigated the frequencies of CMVS in a very large cohort of individuals referred for genetic testing for various disorders.

35:05
And in this study, almost 10% of individuals who received a clinically significant result had ACMV reported as part of that that result.

35:16
But what was really interesting is that is that very small CMVS accounted for a high percentage of the pathogenic CMVS reported.

35:24
So 25% of the the pathogenic deletions or duplications were single or or partial Exxon.

35:34
So these very small CMVS are a significant contributor to genetic disease and that's backed up by our internal data as well.

35:44
But the difficulty is that these small CMVS can be hard to detect.

35:54
There are a number of different ways of detecting CMVS, but detecting CMVS from NGS data is really the gold standard these days because of the high resolution.

36:05
So how is it done?

36:07
At Blueprint we use a method called Read depth mapping, which is where you compare the number of sequencing reads you observe in a region with the number of reads you would expect to see in that region.

36:18
And then if there is a significant difference between the two, this generates a CMV call.

36:25
To detect the MVS from NGS data.

36:28
We use three complementary bioinformatic tools, one which calls deletions and duplications using some software called CMV Kit.

36:38
And then we have an in house algorithm which is specifically optimized to detect smaller one or two Exxon deletions because these can be amongst the most challenging to detect.

36:48
And then we also have a breakpoint detection method that calls breakpoints in the NGS data and these these can alert you to the presence of a CMV there.

36:58
So a number of different methods used in parallel to really try to optimize the detection of different types of CMVS of different sizes.

37:10
And what's really important to note is that for these methods to work well, deep and uniform sequencing coverage is really key.

37:17
So the importance of coverage coming back again, again as discussed in the last case that we went through.

37:27
So in this case, a homozygous likely pathogenic single exon deletion in P gap 1 was identified, ending this patient's diagnostic odyssey, confirming a molecular diagnosis of P gap 1 related disease and providing important information for genetic counselling.

37:49
In that there is a 25% risk to any stiblings of the index patient of also being homozygous and affected.

37:57
And this is thanks to high sensitivity CMV analysis on our wares and the BPG wares is validated to detect single Axon deletions with a sensitivity over 99%.

38:16
OK, so let's move on now to the the last but one case.

38:21
And in this case, we're going to talk about getting our priorities straight.

38:27
And this case involved an adult with intellectual disability, short stature, neuropathy and hearing loss and testing elsewhere had been negative.

38:41
On the West analysis, we identified a missense variant in a gene called more C2.

38:48
This variant was predicted to be deleterious by the Insilico tool that we use.

38:54
It's a very rare variant.

38:56
It's absent in Nomad and it wasn't reported in any patients in the literature to our knowledge and nor was it on disease related variation databases and we had not seen it in our lab laboratory previously.

39:16
So not much to say about this variant other than it's very rare and predicted deleterious spine silico tools.

39:22
So it's a variant of uncertain significance.

39:25
But what was interesting about this variant is that it was the number one prioritized variant by our variant prioritization tool Alvar.

39:37
So Alvar is our in house developed variant prioritization tool.

39:42
What it does is it gives each variant a pathogenicity prediction score between zero and one, with 0 being low probability of pathogenicity and one being higher probability.

39:55
It's based on very many different variant properties over 150 and it works on single nucleotide variants and and short in Dells.

40:07
And it is important to note that it prioritizes variants for us.

40:11
It does not classify them so.

40:13
So it brings the genetically strong variants to our attention 1st and then and then we geneticists interpret and classify the variants as normal.

40:27
So in a W we can use Alva prioritization to rank the heterozygous variants by Alvar score and prioritize those that have the highest score when we do the analysis.

40:43
So those that are the most strong in terms of variant properties, we we look at those first.

40:49
So this, this, this results in an ability to focus in quicker on the likely most relevant variants.

40:57
So making the most of the geneticist time and expertise and a validation in house showed that using ALVAR to prioritize variance in in wears does not jeopardize the clinical sensitivity of our wears, but it does give us greater efficiency and focus.

41:16
So back to our case with the more C2 variant, Morsi 2 is actually a bit of a tricky gene in terms of phenotype matching because because the associated phenotype is is quite variable and broad.

41:29
So there are many different clinical features associated and perhaps a bit of a spectrum to this disease.

41:38
So on OMIM, there are two phenotypes annotated annotated there for this gene.

41:44
So axonal shark at Murray tooth disease type 2Z and then a more early onset syndromic disorder, which is characterized by developmental delay, impaired growth, dysmorphic features and axonal neuropathy.

42:01
So comparing the phenotypes in our patient with those reported in association with more C2 related disorders, we we do have some some phenotype overlap as shown in red.

42:15
So, so our variant prioritization tool alerted us to this variant straight away and then we determined that there was some phenotype overlap.

42:25
So the variant was relevant and worth reporting and the variant was reported as a variant of uncertain significance and then subsequent parental testing did not detect the variant in the unaffected parents, which is consistent with de Novo occurrence in this patient.

42:44
So as a result, this variant was upgraded to likely pathogenic.

42:52
So whole exome sequencing and and subsequent parental testing in this patient identified a de Novo likely pathogenic variant in more C2 and this confirmed a diagnosis of autosomal dominant more C2 related disease.

43:07
And this was thanks to intelligent variant prioritization.

43:12
So our tool prioritized this variant.

43:14
So we were able to focus in on it straight away and we we saw it in the analysis a lot more quickly than we would without this tool.

43:24
So we would have seen the variant anyway without the variant prioritization tool and it would still have been reported.

43:30
It would just take and have taken us longer to to focus in on it.

43:35
But with our tool, this variant jumped out straight away as the number one prioritize variant and indeed it ended up being the diagnostic variant.

43:49
So let's move on now to the last case in this session where we are thinking genotype first.

43:57
And this case involved a baby with echogenic kidneys and liver and poor feeding and poor growth.

44:08
This final case is a demonstration of our genotype first approach to interpretation, which we believe can be very powerful, especially for patients with atypical phenotypes and previously inconclusive genetic testing.

44:24
So what does this mean?

44:25
There are two schools of thought in terms of how to approach exome analysis.

44:31
First of all, a phenotype first approach is where the data is filtered according to sets of genes thought to be related to the patient's phenotype.

44:40
With this approach, you would only see variants in genes that fits with the patient's reported phenotype when you're doing the the interpretation.

44:53
But a genotype first approach is variant driven rather than phenotype driven.

44:57
So this means that the analysis includes variants in all the genes sequenced.

45:04
No genes are removed from the sequencing or the bioinformatic analysis based on lack of perceived phenotype match.

45:12
Rather, the variants are prioritized according to the properties of the variant itself, and then we examine whether there could be some association with the patient's phenotype for those genetically strong variants.

45:27
So our interpretation approach is genotype first.

45:32
So this means that we report genetically strong variants if there is an association with the phenotype, even if the phenotype match doesn't seem to be perfect.

45:44
And that's because patients that come for whole exome sequencing do often have atypical phenotypes or they have rare diseases with maybe an uncharacterized clinical spectrum.

45:55
So we shouldn't necessarily expect a perfect phenotype match in patients that come for whole exome sequencing.

46:05
So This is why we feel the genotype first approach is very powerful.

46:08
But nevertheless, it's also worth mentioning that we do not report anything that doesn't fit at all with the patient's phenotype, even if the variant is a known pathogenic variant.

46:19
So we would not report anything purely predictive, for example, unless it's in the context of the ACMG secondary findings list and the patient has consented for that.

46:33
So we follow a genotype first approach and our reporting is phenotype driven.

46:40
And in the West analysis in in our case, in the baby with echogenic kidneys and liver, we identified a homozygous missense variant in the GLB 1 gene and bialelic pathogenic variants in GLB 1 cause autosomal recessive GM1 gangliocytosis.

47:04
This is a lysosomal storage disease and there is a phenotypic spectrum to this disease and there is an infantel form that typically shows developmental delay and regression.

47:16
Also hepatosplenomegaline cardiomyopathy and skeletal abnormalities, also other characteristic features including congenital dermal melanocytosis.

47:33
So we don't have an obvious phenotype match here with our very young patient, but I said our analysis is driven by the genotype first.

47:46
So let's take a look at the genotype.

47:49
The variant itself is very strong.

47:51
It's rare.

47:52
It's absent in Nomad.

47:54
The variant was predicted to be deleterious by the Insilico tool that we use.

48:00
The amino acid is highly conserved and the variant was not found in the medical literature and it was on Clinvar, but just as a as a variant of uncertain significance.

48:12
But but interestingly, we had seen the variant previously at at Blueprint Genetics in another patient with whose phenotype was very consistent with GLB 1 related disease.

48:24
And that patient also had a second variant in GLB 1.

48:29
And also variants in adjacent codons have been reported in association with with GLB 1 related disease.

48:36
So the point is this variant has multiple very compelling properties and so it came straight to our attention in the analysis and was looked at very carefully, even though the phenotype didn't immediately jump out as a good match.

48:53
So the genotype grabbed our attention.

48:55
So now we have a really close look at the phenotype to decide whether this variant could be relevant for our patient and whether it should be reported.

49:06
So a careful look at the literature did find some cases in the literature where prenatal imaging had shown echogenic kidneys and liver in fetal cases proposed to have this disease.

49:20
There wasn't a lot of information about this because the disease is not typically looked for prenatally and obviously echogenicity is is a very non specific finding, but that was interesting nonetheless.

49:34
Poor feeding and poor growth is a known feature of this disorder, but again it's very non specific.

49:41
But there are a few interesting links with the phenotype here.

49:45
But this was difficult.

49:46
So we had a discussion about this within the team and it was decided that based on the variant being very strong homozygous and some possible links with the phenotype, it was decided that the variant should be reported.

49:59
And it was reported as a variant of uncertain significance because even though the variant is strong, there isn't enough evidence to classify it as anything other than a VUS.

50:12
And then at a later date, we did receive an updated phenotype for this patient and and there were now multiple additional features consistent with GLB 1 related disease including congenital dermal melanocytosis, which is one of the characteristic features.

50:29
And some biochemical studies had found reduced beta galactosidase activity in the patient, which supports a deleterious role on protein function for this variant.

50:40
So based on the now good phenotype match the biochemical studies and all the additional information, the variant was upgraded to likely pathogenic.

50:51
So the outcome of this case was that the family received a molecular diagnosis for this, this this baby of GM1 gangliocytosis, which unfortunately has a very poor prognosis.

51:03
But the family now have the important information that there's a 25% risk of of recurrence.

51:08
And this diagnosis was identified, I think, thanks to a genotype first approach to the analysis.

51:17
I'm not sure if this variant would have been picked up if the analysis filtered based on the phenotype originally given.

51:26
So I think this case is a good illustration of how a genotype first approach can be helpful for patients with atypical presentations or for example, where there's not a lot of information available on the phenotype yet if the patients are very young.

51:47
OK, So I've shown you 6 cases today and six ways in which the limitations and challenges of ways can be met.

51:57
So first of all, through the addition of custom targeted non coding variants on the assay, inclusion of the mitochondrial genome, designing the assay to achieve high and uniform coverage to improve detection of variants in difficult sequence regions.

52:14
And also this high and uniform coverage helps a lot with small CMV detection.

52:20
And then on the interpretation side, intelligent variant prioritisation and a genotype first approach are essential in dealing with the data efficiently and getting straight to the relevant variants for the patient.

52:36
OK, thanks to you all for listening.

52:39
And thanks also to the many colleagues at Blueprint Genetics who contributed to solving these diagnostic odysseys that I presented today, and who are all driven every day to find answers for patients and help families affected with genetic disease.