33 – Functional assessment of missense variants of SYNGAP1

Event Time

May 13, 2021

Here are the introductory comments: 

Our talk for today is “Functional Assessment of Missense Variants of SYNGAP1.”

I have the pleasure to introduce today’s speaker professor Kurt Haas.

He is a professor of cellular and physiological sciences at the faculty of medicine at the University of British Columbia.

The primary goal of Dr. Haas’s research is to understand how brain circuits form during early development and how mutations and errors give rise to dysfunctional circuits that underlie common neurodevelopmental disorders such as Autism Spectrum Disorders and Epilepsy. 

The distinctive feature of his research approach is the development of imaging tools that combine electrophysiology, genetics and microscope design to see brain neural network activity and growth.

Webinar Overview

Dr. Kurt Haas is a professor of cellular and physiological sciences at the University of British Columbia and does research at the Centre for Brain Health. His objective is to understand the genetic architecture of autism spectrum disorder (ASD); however, there are thousands of single-nucleotide variants found in ASD genes and most are variants of unknown significance (VUS), making it very difficult to understand.

In his lab, Dr. Haas has identified 57 SYNGAP1 variants and has developed 7 assays to test their structure function. He explains how the data from these assays can create a multi-parametric functional impact score, an unbiased score that takes into account each of the assays with equal weighting. This score showed varying levels of dysfunction for each variant, suggesting that there are multiple mechanisms of dysfunction – SYNGAP1 dysfunction is not solely caused by haploinsufficiency. He closes with goals for the future, including finding more SYNGAP1 variants, as well as clinically assessing individuals with variants of distinct molecular mechanisms of dysfunction.

Other Relevant Publications by Dr. Haas

Multi-parametric analysis of 57 SYNGAP1 variants reveal impacts on GTPase signaling, localization, and protein stability

Comprehensive Imaging of Sensory-Evoked Activity of Entire Neurons Within the Awake Developing Brain Using Ultrafast AOD-Based Random-Access Two-Photon Microscopy




0:06Hello everyone and welcome to today’s session. My  name is Marta Dahiya. I’m a Syngap parent and director  

0:11of the SynGAP Research Fund. We are very excited  to continue the SRF’s Syngap Research Fund webinar  

0:19series. The goal of the series is to empower your  communications with clinicians as you get  

0:26more clear knowledge of Syngap. We also want to  give you a plug to our next presentation: “Types  

0:32of seizures and EEG patterns in SYNGAP1″ which  will take place on May 27 at 11 am eastern time  

0:43with Dr Ángel Aledo-Serrano. Our talk today is  “Functional assessment of missense variants of  

0:51SYNGAP1″. I have the pleasure to introduce today  speakers Professor Kurt Haas. He is a professor of  

1:00cellular and physiological science at the faculty  of medicine at the University of British Columbia  

1:07the primary goal of Dr Haas’s research is  to understand how brain circuits form  

1:14during early development and how mutations and  errors gives rise to dysfunctional circuits that  

1:22underlie common neurodevelopmental disorders  such as autism spectrum disorders and epilepsy.  

1:30The extreme tech feature of this research approach  is the development of image tools that combine  

1:38electrophysiology genetics and microscope designed  to see brain neuronal network activity and growth.  

1:47A recorded version of this webinar will be  available on the SRF website under webinars  

1:54on the family menu. By the end of the presentation  you will have the opportunity to get the answer  

2:01to your questions. We love to hear from you. Please  write your questions in the Q&A. For those of you  

2:09just joined us welcome, and again our speaker is  Professor Kurt Haas and his talk is “Functional  

2:18assessment of missense variants of SYNGAP1″.  It is now my pleasure to turn things over to  

2:25Professor Haas. Welcome Professor. Thanks Marta. Now  thanks for this invitation to you to speak today.  Outline

2:33The first thing I’m going to do is change my title  a little bit and this is the new title which is an  

2:39email I received from from Hans Schlecht which  is, you know, I think really puts the work in   in a better context which is “Are SynGAP missense  mutations simply haploinsufficient?” and I’d like  

2:51to I think throughout the talk I’ll be sort  of really just coming back to this question.   So my interest is to understand the genetic  architecture of Autism Spectrum Disorder  Genetic Architecture of Autism Spectrum Disorder (ASD)

3:01and other neural developmental disorders and you  know the search for genes associated with  

3:08these disorders was conducted you know has a long  history and the early sort of “low hanging fruit” were  

3:16were discoveries of large either gene duplications  or large mutations that clearly knocked out  

3:22protein function however things have become much  more complicated since it’s been realized that  

3:29the most common mutations are de novo single  nucleotide mutations and that they occur in  

3:39you know hundreds of genes that are associated  with ASD so there are really thousands of these  

3:45mutations now these mutations they cause single  nucleotide changes based in the base code of DNA  

3:52which can give rise to three different types  of mutations nonsense mutations where there’s  

3:58a stop codon put inserted so the protein is  truncated frame shifts where you have a micro  

4:04deletion or addition so all the the codon  is frame shifted downstream of that point so  

4:10that the protein is sort of a nonsense protein  or the most common are are missense mutations  

4:15where a single nucleotide change causes a single  change in the amino acid within the protein those  

4:22small changes are very hard to understand how they  impact protein function so most of them are termed  

4:30vases or variants of unknown significance but my  lab is interested in trying to really understand  

4:36what they do so our strategy which I think  is somewhat unique to our lab is to develop  

4:42large platforms that are that use multiple  assays to study individual proteins  

4:50for what we call deep phenotypic profiling  to understand what individual mutations do  

4:56to very complex aspects of a protein function  proteins are complex little machines they they  

5:04do often do multiple different functions  and if there’s a mutation in their protein  

5:09structure it’s not obvious or clear that you  what function they might be interfering with  

5:14so we think you really need multiple assays  in order to really understand what’s going on  

5:20our goals are multi-fold one is to discriminate  variants that are not likely pathogenic from  

5:28those that are just because we find a variant  in an individual with autism or other disorders  

5:34doesn’t mean that it’s causal so by conducting  multiple assays and finding that the variant  

5:39doesn’t really disrupt protein function  we think that adds evidence that it’s   likely benign and non-causal but other ones  likely are and for one of the ones that are  

5:50it’s a very important to understand what the  molecular mechanism of dysfunction is we need  

5:57to do that at the at the protein level and then  the protein sort of pathway level and then once  

6:04we identify those we can take a subset of those  to move forward to lower throughput assay such as  

6:11for testing for pathophysiology this is  a understanding what’s happening at the   neuronal and neural circuit and whole body  level and eventually this will allow us to  

6:19develop therapeutics having a wealth of knowledge  about point mutations throughout a protein also  

6:26provides invaluable structure function analysis  or information about what proteins really do so  

6:34ultimately our goal is to link gene mutations to  molecular dysfunction and pathophysiology and all  

6:40these to disease expression and therapeutic  outcomes and and together this is really the   foundation for personalized medicine and this  sort of approaches is sort of diagrammed here  Functional Variomics – Personalized Medicine

6:49and here you have individuals with autism  spectrum disorder or other neurodevelopmental  

6:55disorders and they usually present in a spectrum  with very complex and very a range of different  

7:01phenotypes and if you’re interested in  understanding what the cause might be  

7:07you know the strategy is to you know then sequence  these you know the genomes find mutations and  

7:15at this point largely we’re stuck with relying on  predictive algorithms to determine whether these  

7:21variants are deleterious or not and while there’s  been a really impressive power in these algorithms  

7:28they’re not perfect and they should not be a  substitute for for real experimentation so this   is where we come in and test these variants at  the protein level and the protein sort of circuit  

7:40level and this is data can then be fed back to  help develop or improve predictive algorithms  

7:48and once we’ve used these methods for  high throughput analysis of many variants   we can find sort of representative variants  that are representative classes of dysfunction  

7:58which can then be brought forward to these sort  of slower throughput assays for understanding  

8:04pathophysiology what’s happening at the cellular  or animal level and then all these sort of  

8:10this information at these different levels are  useful for developing targeted therapeutics that  

8:17are specific to specific molecular dysfunctions  specific pathophysiologies and then feedback  

8:23to the to individuals so this is where I’ll be  talking about today this sort of level of just  Molecular mechanisms of missense mutation dysfunction

8:30dealing with variance protein level and sort of  the the protein sort of a molecular pathway level  

8:38and we’re trying to understand what are the  classes of the protein dysfunction and there’s   we talked about haploinsufficiency typically  that means the the protein is just absent and  

8:47it’s just sort of gone but there’s we often talk  about loss of function and this could be partial  

8:53loss of function or complete loss of function  but to be clear that function is under whatever  

8:58you’re calling as the function so there are  many functions of proteins there’s also gain of   function and dominant negativity when we look at  what proteins actually do in love for dysfunction  

9:09this is a short maybe not even complete list of  all the possibilities of what might be happening  

9:15there can be impacts of seeing of mutations  on protein stability subcellular localization  

9:23uh post-translational modification so  phosphorylation my restylation pulmonation  

9:28there’s tremendous amount of these  modifications that regulate protein function   and they could be disrupted many proteins dimerize  they bind to themselves so that could be disrupted  

9:38but certainly most proteins bind to other proteins  so there’s a lot of protein protein interactions   if they’re enzymes then there’s both  catalytic activity and substrate interactions  

9:48so any of these processes can be impacted  it could be one or a combination of themMethods

9:55so these are the methods that we use in my lab  for high throughput assessment of variant function  

10:01we construct large dna libraries of individual  genes between ten you know tens to hundreds of  

10:09of point mutations we physically make them  and then we put them into expression vectors  

10:15which are plasmids we can put into cells and  they’ll be expressed and we put them into   these sort of genetic tools and this one is  an important one we’ve developed which drives  

10:26two proteins off of the same promoter so it  uses this P2A internal it’s a self-cleaving  

10:32peptide from viruses what allows us this is the  pro the promoter that drives first expression  

10:37the first gene which is a red fluorescent  protein rfp it goes reads through the P2a  

10:44and then that self cleaves but then makes another  a copy of the downstream protein which is a green  

10:50tagged you know protein variant whatever gene of  interest is here’s PTEN so that’ll come out as  

10:56a fusion with gfp and this is very powerful  because it allows the red to be used as a  

11:02protein expression level and then the  gfp tag to show us where the protein is   or how much the protein if it’s degraded and then  we express these vectors into human cells such as  

11:17human uh embryonic kidney cells hex cells these  are really workhorses cell lines in labs that can  

11:24be grown whatever you want to easily transfect  it and then then use for experimentation for   for expression and this is the the readout  we use fluorescence imaging and that’s either  

11:37with high content imaging or flow cytometry  and we’re detecting either the fluorescence   of this red and green fluorophores or applying  antibodies that detect typically phosphorylation  

11:49states of downstream protein markers so if  proteins are typically in molecular pathways  

11:55where they can then activate or suppress inhibit  downstream proteins and that can be a marker of  

12:01the protein’s function and by having antibodies to  the activation state of those downstream proteins  

12:07or markers we can detect protein function  so so we’ll we’ll have fluorescently tagged  

12:14antibodies and we’ll look at them in individual  cells so on the high content imaging we are  

12:22imaging these fluorescent hex cells and this is  done in an automated fashion very high throughput  

12:28and then there’s computer tools to segment these  cells and detect fluorescence within them and  

12:34make large measurements so you can really  move through large numbers of cells at once   and then the other tool that we use a  lot is called flow cytometry here these  

12:43cells are being placed into a solution and then  they’re passed one at a time through a detector  

12:50lasers are shot at them and the fluorescence is  detected so we have single cell resolution of  

12:55all the markers that we’re interested in i’m going  to start by not talking about SYNGAP1 but talking  

13:02about a different protein PTEN that i think gives  us a good idea of what our approaches and what  

13:07one can learn and then i’ll move to our sync up  one studies so for PTEN we recently published a  

13:13paper last year where we studied 127 variants in  18 assays in five model systems so this is really  

13:21deep phenotypic profiling and now we’re actually  up to 21 assays in something on the order of 150  

13:26variants p10 is a really well characterized cancer  in autism associated gene and it acts to pretty  

13:34much inhibit this protein here AKT. AKT promotes  cell growth and proliferation so by inhibiting it  

13:43that’s how PTEN inhibits cancer and it mutations  to be 10 cause it causes dysfunction or overgrowth  

13:51in neurons leading to two autism issues so what  we’ve done is down here is the p10 gene with the  

13:59different domains and all these lollipop figures  are are are in missense mutations that we’ve made  

14:07individually on different copies of the of the  gene and they’re color-coded where the blues are   found in individuals with autism spectrum disorder  or intellectual disability so that’s this group  

14:17the blue group and you can see there they’re found  throughout the protein and the other important   groups would be these biochemical controls these  are ones that we know interfere with function  

14:28and we also include these purple ones which  are variants found in the normal population  Protein Stability Assay – Flow Cytometry

14:35so and this is what the data looks like this is a  flow cytometry sort of data set each point here is  

14:42the response from a single cell you see there’s  there’s tens of thousands of cells this actually   results from three experiments laid on top of each  other from these three different variants of PTEN  

14:51wild type a hyper stable one and a unstable  version and what we’re worth plotting is the red  

15:00fluorescence and the green fluorescence for each  cell again the red fluorescence says how much of   the protein the varying protein was made the green  reflects how much still remains as the protein  

15:11gets turned over so taking the green to red ratio  tells us what the stability of that protein is and  

15:18that’s shown here in those the right the sort  of medium purple is the wild type this is the  

15:24plotting the gfp the green over the  red and this is sort of frequency from   these sort of plots and this salmon color is a  destabilized variant that shifted to the left  

15:35the dark purples as hyperstable has one shifted  to the right so so that’s the data set and  

15:41this is the data from 127 variants so this is  protein stability it this data is all normalized  

15:49to wild type so this is the wild type PTEN and  i’ll just have you focus on first this group here  

15:55which are the blues these are the variants from  autism or intellectual disability and the results  

16:00were pretty striking this is the first stream we  really looked at and i was sort of expecting just   complete instability or wild type stability and  that’s not what we saw we found variants that were  

16:12wild type levels some were hyper stabilized and  then there was this full range of instability  

16:19again all these are induced by single point  missense mutations so that was really dramatic  

16:25and this really gave you know sort of support  to maybe this haploid insufficiency model that  

16:31the disorder is really due to gradual depending on  the case a gradual decrement in how much PTEN is  

16:38present and that that sort of fits interestingly  if we look at these purple ones over here these   are the population variants many of them are wild  type but some are a little bit hyper stabilized  

16:49and a couple are destabilized showing  that even in what’s considered the normal   population for these databases there  is protein variance in protein function  

17:01okay the next thing we we did is to look  at function by looking at the activity of  

17:06this downstream protein AKT using antibodies  against the phosphorylated state of phosphate  

17:13and that’s showing the flow responses again the  the medium purple is wild type in this case the  

17:20c124s is a is a known non-functional PTEN so  that’s shifted to the right so there’s more  

17:28fossil AKT and a gain of function for a shifts  it the other direction and this is the data from  

17:36large numbers of variants and again if you look  at these blue ones the autism associated variants  

17:42this was really striking this shows you the  function of these variants and again it’s a  

17:48very complex profile it’s not all or none there’s  a gain of function that somebody to this one here  

17:55which is this known gain of functions for a here  again all this data is normalized to wild type is  

18:01one and the empty vector gfp by itself is zero so  so we have a gain of function we have a few that  

18:08are wild type like and then there’s this range of  partial loss of function complete loss of function  

18:15and then these which are dominant negatives so  they’re worse than having an empty vector so this  

18:22is really important only five dominant negatives  were known before this we found 27 additional ones  

18:29so this suggests that there’s a very comp there’s  a large complexity in the molecular mechanisms of   dysfunction for this gene and what was important  is when we went back and compared the protein  

18:42stability data to the functional data that’s  plotted here where stability is in the x-axis  

18:48protein function is a readout for the downstream  levels of its ability to inhibit phosphate  

18:57you see that we found that there’s two groups of  of variants each point here is a different variant  

19:02there are those the blue ones where the function  the amount of function or dysfunction tracks  

19:10directly with how much stability or instability  there is within the protein so for these i  

19:16think that the the molecular mechanism is haploid  insufficiency so it’s just an absence of the PTEN  

19:23however the red ones the amount of  dysfunction is not correlated to the amount of  

19:30of stability so a different mechanism is going  on and if we look at these red or green variant  

19:37populations along the the length of the protein  we see that the red ones fall into specific  

19:44domains and these are known substrate binding  and catalytic catalytic domains so it really  

19:50says that yeah if there’s a variant within  these domains it’s not going to be haploid   sufficiency that’s that’s causing dysfunction it’s  because of a disruption of of enzymatic function  

20:01okay so that’s i think an important sort of sort  of perspective in general we next move to making  SYNGAP1 mutations and disease

20:08to studying SYNGAP1 clearly there’s many single  nucleotide variants of SYNGAP1 being discovered  

20:18that are these include nonsense  missense and friendship mutations   and it’s associated as you well known with idsd  epilepsy and schizophrenia but even recent reviews  

20:30such as this one this is a direct quote from this  paper in 2019 talks about the novel variants of  

20:38Syngap resulting in haploinsufficiency lead to  a defined phenotype and that this may explain  

20:45up to one percent of ID cases and i just i’m a  little more cautious about making this this sort  

20:51of statement i wonder whether other mechanisms  besides haploinsufficiency might be happening  

20:59so this is a diagram from a review by clement  just showing you some of the molecular pathways  

21:06associated with SYNGAP1 SYNGAP1 the gap is  it’s you know it’s a synaptic gap it’s a  

21:12gtbs regulating protein and it’s known to  regulate both ras and wrap these gtp aces  

21:20uh you know these pathways molecular pathways are  very complex when you start putting gt bases in  

21:26them things get very very complex but so this  is a really a subset of a very small subset  

21:32of what’s going on but you can see there are  downstream proteins such as erc and p38 map kinase  

21:40and then downstream of both of those is kreb  so and i won’t go through this sort of function  

21:48like what what SYNGAP1 is doing because you  probably are well aware of this but it’s   associated with both neuronal growth and with  synaptic plasticity uh very well connected to  

21:59ltp you know lots of work from Rick Huganir’s  lab but it’s also connected as well to ltd so  

22:05it’s already in this balance between the two  where the Gas pathway is more connected to LTP  

22:10and maybe the rap is more connected to ltd  so we wanted to look at markers downstream  

22:17of these pathways so we picked the reds are where  we have antibodies against phosphorylation state  

22:22detecting the activation of these downstream  proteins so we have markers for activated erc  

22:29and p38 map kinase as well as a much more  downstream transcription factor crab we also  

22:38did a quick screen for additional proteins that  we know are downstream of Ras and we found that  

22:44gsk3 beta is also regulated by SYNGAP1 so  we’ve included that in our screen as well  

22:52so uh just to quickly review the structure of  SynGAP it has multiple domains including the  

23:00uh these plextron homology or ph domains that  are likely involved with membrane recruitment  

23:06there’s c2 in the gap domain which is important  for the the gap gdb it’s activation activity  

23:13and then there’s this large c terminal domain  that’s uh called a structurally disoriented or  

23:19disordered domain less is known about this  it’s only it’s known that there is these  

23:24sh3 and a coiled coiled domains that are  involved with the protein protein interactions so  

23:32that’s the structure and these are the variants  that we we’ve identified we found 57 variants of  57 SYNGAP1 Variants – 7 Assays

23:38SYNGAP1 that we’ve included in our study and we’ve  developed seven assays to to test their structure  

23:45function these include a wild type version 23  that are in individuals with ASD or ID I call  

23:52these unconfirmed because again just because  they’re identified in these individuals do not   mean they’re causal even though they were calling  them this we don’t think that they necessarily are  

24:01causal we also included 16 population variants  these are from the nomad database so this is  

24:08are thought to be non-disease associated sort  of normal population controls and those are  

24:15the the blues are the the asd/id variants you  can see there they’re found across the protein  

24:21and we picked the green ones the population  controls as being across the protein as well  

24:27and then we included 17 what we call biochemical  controls these we think are likely to be loss of  

24:32function gain of function either you know it’s  it’s known like most proteins that SYNGAP1 is  

24:38phosphorylated is probably more than 20 sites and  these are the kinases that phosphorylate them so  

24:45we picked sites throughout us Syngap that are  the sites for phosphorylation we mutated them  

24:54as well as loss of function variants that were  studied in non-human orthologs so we we include as  

25:02many as we could some we just developed on our own  hoping that they might be non-functional and this  SYNGAP1: Stability Assay

25:08is the data so this data is presented differently  than the data i showed you with a p10 here we’re  

25:14not normalizing the data so i’m always going to  show this is just pure stability the black is  

25:20the wild type stability and uh so it’s black red  is gfp that’s not very meaningful in stability  

25:26sense here so don’t but it is more informative  in other plots and you can see the spread of the  

25:34diff of the 57 variants that we tested again  the the blues are the autism id associated  

25:42variants the greens are the population controls  and the purples are the biochemical controls  

25:49and you can see that there’s uh that in general  SYNGAP1 is resistant or largely resistant  

25:56to missense induced destabilization  and this is likely because it is such  

26:02a large protein it’s like a thousand  three hundred you know 43 amino acids  

26:10dead there are studies that show that longer  proteins are less sensitive to instability  

26:15though the mechanism for that isn’t completely  clear but that seems to be true here so there are  

26:22only three missense variants here that show  significant instability the two these two are  

26:32nonsense early stops at 687 579 they show strong  instability but all nonsense mutations should  

26:41not be considered to be destabilized there’s  another one right here that is indistinguisha  

26:46it’s not significantly different from wild type it  actually shows about 88 of wild type stability so  

26:53this is likely present so i wouldn’t say that that  nonsense mutations have to be to fall into this uh  

27:00you know haploid sufficient category they’re still  present and they might still have partial function

27:10and then there are these three over here  that are hyper stable two autism and one   bio chemical control so hyper stability is another  mechanism and when we look at where the very rare  

27:23ones are either hyper stable or unstable they fall  within these sort of discrete domains within the  

27:29gap domain or the c tail region next we looked  at function so this is looking at uh antibodies  SYNGAP1: ERK

27:36against phospho erc and again this data is not  uh is not normalized the block is the wild type  

27:44the the red is the empty vector just gfp so  this would be loss of function all the black uh  

27:50stars represent different significant difference  from the wild type the red ones are significant   difference from the gfp so what we find is that  wild wild-type uh Syngap reduces phosphorylation  

28:04of arc by 28 and there’s complete loss of  function in 18 of the 57 partial and 22 of them  

28:14so and the the lowest ones here are are sort of  biochemical controls and as well as the highest  

28:22ones so the highest ones are more functional  than uh or they’re equally functional wild type  

28:28okay so when we if we plot out certain  irk function across the protein structure  

28:36it’s very clear that there’s a distribution  so what’s what’s shown here is the gfp value  

28:42that’s the line here and the wild type  is this black line so variants that are   closer to the black line are more functional  those are closer to jp are less functional  

28:50and you can see that that ones near the n-terminus  none less than amino acid 4 85 show less than 50  

29:03function whereas those that are above 800  all of them show significant dysfunction

29:11the next assay we look at is gsk3 beta  this these results correlated well with  

29:21the data from irk you see this  and so there’s a is a you know  

29:27significant correlation between these two markers  so this is important because this shows a novel  

29:34interaction with a new sort of downstream  pathway or Ras pathway of this important protein  

29:41gs k3 beta is important in both in synaptic  plasticity and neuronal growth as well

29:49then there’s the other pathway that the sort of  rap ltd pathway for the marker p38 map kinase  

29:58we find less but certainly significant  impact we find the results are correlated  

30:05somewhat with the irk results but there’s  a lot of variants that are either show  

30:10more function in one pathway versus the other  and this raises the possibility that different  

30:17variants are selectively inactivating an  ltp pathway versus an ltd pathway and so  

30:24it’s important to understand what that might  result in at the path of physiological level  

30:30and then finally crab which is this  downstream transcription factor  

30:36we found that while there was some dysfunction  we find that and that all variants were able  

30:47to uh inhibit uh crab phosphorylation compared to  wild type compared to gfp so you see that they’re  

30:54all show some activity but some so partial  loss of function and then there is some gain  

31:00of function phenotypes over here it’s interesting  that even in cases where uh irk variants are don’t  

31:09suppress irk phosphorylation they all can affect  this downstream crab suggesting that maybe it’s  

31:20working through different path multiple pathways i  won’t talk much about this next data set but it’s  

31:26we’ve also looked for sub cellular localization  of SYNGAP1 of course SYNGAP1 is very important  

31:32because it localizes to synapses and you know one  can make the argument that well what’s the utility  

31:38for studying this in a non-neural system we think  the pathways are still intact so that’s important  

31:43but when we looked for subcellular localization  we also found a phenotype so on top are these  

31:49are hex cells just expressing gfp and you  can see the gfp throughout the cytoplasm   but when we express wild-type SynGAP we find  that it’s it localizes to puncta in the cytoplasm  

32:02and when we look at different variants we  find that there’s either a decrease or an   increase in the size of these puncta so it raises  well a question of whether this is important  

32:13for its function in neurons this is the mean  speckle size or the puncta size this is the  

32:20wild type you see that most are similar though  they are significantly larger than wild type  

32:26some are slightly smaller and then there’s a  subgroup that are all significantly much larger  

32:33than wild type and these fall into the  the two early stop variants we studied  

32:40as well as two variants that are very unstable and  three additional ones that are very close to the  

32:46the c2 domain so something very interesting  is happening here with these variants and  

32:52this shows where they’re located across  the protein so when we have these large   data set with seven different assays we can then  do clustering analysis to see which variants are  SYNGAP1: Clustering variants by functional responses

33:02similar to each other to try to see if there’s  uniform molecular mechanisms and this shows  

33:09an example where here we have one that’s  this r 596a which is a biochemical control is  

33:16very different from every other  variant so it’s important to find those   and then we can find ones that are very similar so  these two here which are r485a and m759v and this  

33:28other one this other group here these are the most  similar to each other throughout the the entire  

33:34group and what’s important is that even though  they’re so similar in their phenotypes they’re  

33:40very far apart from each other in the protein so  you know of course proteins are three-dimensional  

33:47have three-dimensional tertiary structures so it’s  important to identify these to try to understand  

33:52why they’re giving such  similar molecular phenotypes   from this type of clustering we can identify  variance we can reduce this data set to a smaller  

34:03number of representative variants that we can  then take forward to lower throughput assays and  

34:09this is also aided by pca or principal component  analysis using multi-dimensional clustering using  

34:15all of our data sets and using this approach we  find five groups of variants and it’s driven by  

34:23these these different features whether they’re  likely loss of function across assays likely  

34:28wild type or other other features so this is a  very important approach for for a reduction of  

34:34the variant class and i’m going to sort of wrap  up by by showing this which is the a heat map  

34:41of the the response of each variant for all of  our assays our seven assays including stability  

34:49as well as a multifunctional score which is  an unbiased just sort of a score that takes in  

34:56into each of these assays with equal weighting so  i think one i could debate the utility of this but  

35:01i think it’s it’s a first pass it’s what we’re  using but when you look at these patterns where  

35:06white is wild type function and blue and  supposedly dark blue is loss of function  

35:12what’s striking is that there’s a mosaic or  speckle pattern across this these variants  

35:18where it’s not just all blue it’s not all  or nothing throughout different assays  

35:24and that’s highlighted by these these dots that  show dysfunction in these these these with these  

35:30markers and it’s very you know important where  you take one like this here where you have strong  

35:37impact or dark blue in irk and gs k3 beta but you  have function in the assay or this one or this one  

35:46or this one the irk and jsk gsk3 beta are fine but  there’s dysfunction in the p38 pathway and this  

35:56one is likely sensitive to crab so this really  highlights or suggests that there’s multiple  

36:03mechanisms of dysfunction that have differential  impacts in their their molecular circuitry  

36:09so when we talk about haploid sufficiency it’s  hard to sort of reconcile that and and is it  

36:14possible to say well is there a function specific  haploid sufficiency that makes sense so i think   a more a deeper more sophisticated approach for  describing these variants is probably important  

36:26when we look at the association of variants  to their different sort of groups again  

36:33this is ASD or ID our biochemical groups  or our population controls and this is the   multi-functional impact score and it’s probably  important to do this sort of analysis with each  SYNGAP1: Disease Associations

36:43of the different assets as well but we’re just  showing you this data if you see this it’s very   clear that that the dysfunctional ones down here  we think are likely causal but there are other  

36:53ones that are very wild-type-like and we think  it’s possible that these might not be causal they  

36:59just have to be found in individuals with ASD or  ID our biochemical controls some of them are very  

37:06impactful it’s very important to identify  these to allow them to be brought forward   for as control very solid controls for studies  of pathophysiology so it’s important to identify  

37:16the ones that are really loss of function and then  it’s important to also realize that in the normal  

37:22supposedly normal population there’s a spread of  protein function so it’s not everything is not  

37:27just one like wild-type level so it’s possible  these people have not yet been identified as  

37:32having disease or there might just have a  change a difference of susceptibility to disease  

37:38and what’s shown here are the the the criter  or the categorization of these variants found  SYNGAP1: Pathogenicity Classification (ClinVAR, Literature)

37:46either in databases like clinvar or the  literature so pre-assigned associations  

37:52as pathogenic or likely pathogenic benign  likely benign thus is a variance of unknown   significance or those that are not scored and  you can see that using our multi-functional  

38:02scoring we find that there’s justification  for some but other ones we think should be   looked at because these ASD ones we don’t  think are likely pathogenic whereas the ones  

38:13declared as being benign we think there’s a lot  of dysfunction here so they should be revisited  

38:18but we can also aid in providing some  uh so maybe uh classification of these  

38:26buses or those that are not yet scored so  i think with that i’ll end with this sort  Conclusions

38:31of conclusion you know going back to this question  are some syngap mutations simply haploinsufficient  

38:38well we find a diversity of impact on downstream  markers in both and, id and the normal populations  

38:45suggest that there are multiple  molecular mechanisms of dysfunction   and importantly this might differentially  impact these these important synaptic pathways  

38:54of ltp versus ltd we would expect that this  might give rise to different pathophysiology  

39:00and importantly we find minimal impact on protein  stability and i think this is the impact of  

39:05protein stability would be the most justification  for using haploinsufficiency as a mechanism  

39:11so for moving forward you know we’ve created this  platform and it’s really we just see this as the   start we can always add new variants and add more  assays so we’re always searching for new variants  

39:23to test and I think we have to expand our assays  to either known or unknown downstream markers  

39:33we need to then take forward our our most  impactful variants our representative variants of  

39:38during classes to studies of pathophysiology  in neural cultures and animal models  

39:44and we really need well detailed clinical  assessment to link these specific molecular  

39:51mechanisms to specific patient phenotypes so  that’s it i’ll just thank my lab and my generous  

39:58funding from the Simons foundation and from the  Canadian institute of health all right thank you

40:09that was great thank you so much that that  really really was very good and yeah and we  Q&A

40:15have the first question is how do you close the  saying that you guys need more variants and and we  

40:23get that question very frequently about getting  new variants and getting more kids tested and  

40:31how are you guys doing that process you know we  research the literature we we work we have close  

40:39contacts with the simon’s foundation and other  foundations but we’re we’re always open to reach  

40:44out to groups like yours to to find these variants  okay but there is a process that you guys follow  

40:53through like blood samples or because we are going  to have no so we don’t we don’t do any of the  

40:59sequencing ourselves we we always get our variants  from other sources and ideally there there’s you  

41:06know good clinical assessment that goes along with  them okay okay yeah we have some collaborations  

41:13very similar yeah that’s one question we  get along and i think hans is here and jenny  

41:20jj jr is here too and and they probably have  more important questions that i have for sure  

41:27hey hans hi can you hear me okay yeah dr haas  that was fantastic i mean you know the paper  

41:34itself is as dense as cheesecake to get through  scientifically so a you know thank you for the for  

41:41the miss sense parents and your team because it  takes a stunning amount of time by everyone lined  

41:48up here for the photo i put some questions in the  in the chat and sort of flipping back for screens  

41:56you know the first one I guess it’s remarkable  that a couple of variants had significantly  

42:04increased protein stability now granted that comes  with a level of dysfunction physiologically but  

42:14could the two be separated you know i’m thinking  about my son who has a nonsense variant where  

42:19potentially promoting protein stability could  you know correct haploinsufficiency by a certain  

42:26percentages hopefully to benefit the phenotype  so i guess that’s that’s my first yeah that’s a  

42:34really important point and you know it’s always  it’s a i think every protein is unique and it’s  

42:41it’s sort of functional range the range  of protein concentration that’s that’s  

42:47necessary and it’s also probably unique and  whether if you have if you have too little or   too much if that matters and yeah i think that’s  something that we have to determine you know  

42:59when we see if we find variants that are hyper  stable i guess we should really look closer to  

43:04see if those are the ones that are having gain  a function in our assays and test those with  

43:11pathophysiology i think that’s what has to be done  and i think with if you have treatments that that  

43:17increase stability i think you know you definitely  want to do a pathophysiology to study to see  

43:23how much to are you increasing it and  what that how that impacts function or  

43:29integers just in the past six months there’s been  you know headlines about alpha fold in terms of  

43:36predicting tertiary structure are and you brought  up tertiary structure is there plans to try to  

43:42correlate you know that genotype phenotype to to  tertiary structure yeah i mean that that’s really  

43:51important and work and and it’s an area that we’re  we are moving into but we’re just developing our  

44:00expertise and we’re our team is we have a lot of  on our plate right now so but i think probably  

44:07just find the right collaborators to work with  to hand our data to so i think having this  

44:12type of deep profiling is very valuable for  modeling for for both bioinformatic you know  

44:19predictive assessments but that type of tertiary  structure yeah so we we have to reach out and  

44:25and make those connections let me say thank  you again and and let uh jr ask her questions  

44:34hi thank you yeah thank you so much professor  haas that was amazing i’m sorry i don’t quite  

44:40understand the stability aspect did you is it a  steady state expression that you’re looking at  

44:47or did you pulse and look at a time after so yeah  it’s steady state i mean it’s like a good point  

44:53so would you call that persistence or you know  you caught your stability i mean abundance would  

45:00probably be a best okay yeah yeah okay  abundance that makes sense so have you  

45:07i don’t think you did this but could you ever  like heat the hek cells a little and actually   look at this you know like stability to me  is more like heat stability that kind of  

45:17that kind of thing so i wonder if  you could just up the temperature   yeah that’s an interesting idea we have considered  you know that there’s complications with i mean  

45:28other things that happen but and there’s only  certain ranges one can do with textiles but   you can definitely change their temperatures and  yeah we probably should get around to doing that  

45:39it’s great that you’re making these  high throughput assays that can really   you know look at look at a lot  of things so thank you for that  

45:47and i think we’re all going to be really super  interested to see the so you showed us some  

45:56graphs about what was it the  what did you call them the  

46:03at the end and you said you only had the  main one but you wanted to do it for each   assay like that kind of thing yeah the multi yeah  the multifunctional score but looking at that for  

46:11each essay i think will be really interesting  and making kind of making a picture of what  

46:18seems to be happening at at different points along  the protest right exactly ideally we’d have very  

46:25nice clinical assessments and  we’d be able to link up specific  

46:30molecular pathway dysfunction to specific  phenotypes that that the the most ideal  

46:36situation i think that’s what the field  needs and hopefully we’re all moving towards  

46:42yeah so when we need it all linked up to  patient data right to see what what’s going on

46:52yeah I guess i’m really i’m just really surprised  at how much the genome may be that you know the  

47:00the population variants were right in there  with with everything else and so that made  

47:05me wonder if really just focusing on the  ends you know the ends after that after the  

47:11so here’s if if green’s sprinkled in there  just like focusing on the part where there   isn’t green if that is going to be more  predictive yeah I don’t know my feeling  

47:22I feel a few different ways about this one one  especially in autism like SYNGAP1 is definitely  

47:28one of these genes where if you have a strong loss  of function you’re very likely to have a problem  

47:34the other gene is probably less so but you know  certainly in autism we think that there’s a lot  

47:39of polygenic combinatorial effects that  go on and that all of us likely carry  

47:46mutations that predispose us for this but it’s  really the combinations and i think underst i   don’t think that they’re insignificant i think  that they’re significant in the right context  

47:57right right well and i’ve i know that when i was  first hearing about Syngap i was hearing that  

48:03less severe so missense mutations can be  associated with bipolar disorder schizophrenia  

48:10and autism without intellectual disability so  i’m wondering at what level and you and you even  

48:17mentioned that you said well maybe people  have an issue that has yet to be diagnosed and  

48:23some of those things are  diagnosed later in life so

48:28yeah yeah that’s a very important point yeah

48:36have i don’t think you can do this high  throughput that i’m going to ask anyway  

48:41so Syngap is a developmental gene and so it’s  going to affect the the structure of the brain  

48:49from the get-go and so to my mind some some  of the variability could be in brain structure  

48:57in in how well parts of the brain are  connected to each other cell morphology  

49:06and then and that’s all you know  maybe pre-birth right and then after  

49:15after a mouse or a human is born then this the  the Syngap variant is gonna make make the brain  

49:26work differently right like like our brains  are experiential we build our brain by what   we experience and so it’s going to be it’s going  to be built you know slightly different that way  

49:37too so i’m just wondering are there any of those  is there anything there that you think you could   look at is there is there’s cell morphology  or that’s just not going to be high throughput  

49:48yeah so this is this what i call pathophysiology  right what’s happening at the real neuronal  

49:54level both both you know synaptic structure  and then circuit and intact organisms and  

50:01we do some of that work in my lab and  we have collaborators that do this in   different model organisms and it’s it’s it’s  not high throughput it’s it’s very laborious  

50:12you know what we what i think is important from  our high throughput work if we can identify study  

50:19you know 60 or 200 variants but we can reduce that  to you know five or ten representative variants  

50:29that are very common to the other ones and then  and move those through our lower throughput assays  

50:35you’ll be maybe informative transferable to the  other variants that are within that class so that  

50:40that’s our our strategy for forward on this in a  way that’s that’s useful yeah well i think well  

50:47we would agree that we’ll miss less you know we’ll  miss less important things if you look at look at  

50:53all these essays so because they’ve already been  somewhat surprising yeah so what would you say for  

51:03us moving forward when we have people who  have a missed sense mutation and we’re not  

51:08sure what it is like what’s do you have any  sort of practical advice for looking at of us

51:17you know obviously there’s predictive  algorithms but you have to know which of those   really fit Syngap so some clearly don’t  for Syngap it’s very strange so i think  

51:29what we’re trying to do is use our data and really  find the predictive algorithms that are best  

51:35eventually in long term is try to improve that  those algorithms based on our our studies you know  

51:44yeah i i don’t know i think you know that’s  as far as therapeutic approaches you know i  

51:53think everyone just has to keep an open mind to  different molecular mechanisms of dysfunction  

52:00yeah but at this point i don’t i i’m feeling  you’re not necessarily maybe you are open to   us like emailing you and saying what do you  think of this one i mean what what’s the  

52:08what are we is that something we can do or is that  something that is coming later and you’re like   yes in two years so i’m gonna have this algorithm  and it’s freely available to geneticists you know  

52:18genetic counselors or i think really practically  we’re very open to receiving new variants ideally  

52:28they’d be put into a database like clinvar so  we can reference them saying this is a reference  

52:34this is the site and it has some clinical  assessment associated with it so that’s  

52:39that that’s the best thing we can handle and  ideally we’d you know do repeat this study  

52:46with more variants and with more assays so  you know like as we as i said with the PTEN  

52:54project we’re we’re now engaging you  know that’s developing even though we did   127 variants in 18 assays we’ve developed more  assays and we’re going to put in probably another  

53:06hundred variants maybe so these things are this  platform is established and now can develop as  

53:14long as we have the people to do it you know it’s  it is labor intensive but that that’s what isis  

53:21is being able to sort of perpetuate this at the  same time working with you know pathophysiology  

53:31studies and helping with  predictive algorithms and all this  

53:36okay thank you so much it’s  a lot of interesting work   one of the things that catch my attention was that  you showed some variants that they were classified  

53:47pathologic but then when you did the function  analysis you you said they are not likely to be  

53:54pathologic they are not likely to be causing  Syngap they i mean the the phenotype then  

54:02how often that happen it does happen it  is happening with our PTEN studies so  

54:08yeah this happens so how variants get classified  as pathogenic or benign sometimes it is not  

54:19very well done i guess other times it’s better  it’s and the sort of the boards that make some  

54:28of these decisions they don’t rely on functional  assessments like laboratory assessments typically  

54:34if there’s a there’s a standard that that the  functional assessment is only one component  

54:40of deciding whether something is it falls into  those categories and that requires you know under  

54:45their guidelines they’ll say well a functional  assessment for like PTEN would only be associated  

54:51with the phospho-akt pathway and nothing else no  other functional assessment matters to them and so  

54:59it’s it’s it’s a very simplified a simple approach  i don’t think takes them to the complexity of   protein protein function in the relationship of  those functions to disease so this is another area  

55:08this classification has to evolve as these studies  go move forward i understand that the when talking  

55:17about pathological predictions a lot of it’s based  on you know clinical assessments which i think is  

55:22is very fair but i think functional laboratory  assessments like what we do has an important  

55:28impact and it you know and that there’s probably a  diversity of functions that should be incorporated  

55:36this made me think that probably in five years or  six years we’re going to have to kind of reject  

55:42the function of the protein of our kids kind of an  individual do you think that’s going to come and  

55:48because it’s a precision medicine question then  i guess you have to be more precise at some point  

55:55do you think that’s coming on the pipe then  yeah yeah this is this is the future and  

56:01it’s things are moving faster and faster the  ability to do more and more high throughput  

56:07studies like this in larger volumes  is it’s becoming faster and cheaper so

56:16are you planning on doing more putting more stop  mutation protein truncating variance through your  

56:22thing just because the three had two two had one  phenotype and one had a different type yeah we we  

56:29have some and i think we have ten more in hand  actually that we will probably look at great

56:37all right so sorry no go ahead please no  things are always surprising we we never  

56:42expected the diversity in the the general  population and we didn’t expect the stop   coding to really show much so we initially  throw in a small number and they’re like oh  

56:50initially with our population controls for their  first study we put in three we thought oh they’re   all gonna be a while time we’re just wasting  our time they all had very diverse responses so  

56:59we’re learning as we move as we go along so i warned you before they got on that  hans jr would beat you up in a good way  

57:11so i’ll give you i’ll give you a softball from  where i sit as a non-scientist who talks to the   two you know the 762 patients we’ve  gotten in the world 215 in the u.s  

57:20there’s a 2019 paper out of invite that  says Syngap was their 10th highest hit   with 39 pathogenics or likely pathogenics but  for those 39 they had 79 buses right so so my  

57:33world is is one of like you’ve got a pathogenic  or a likely pathogenic and then we pull you into   our digital natural history study where by the way  we have over 100 ps and lps with complete medical  

57:43records available for researchers when you want to  tie up patient data and we have a list of variants   that JR’s probably already sent you but on top of  that i have all these buses who are like what do i  

57:54do my kid sounds like a singapi and kids with NDDs  often you know they sound similar but when you   start getting into certain things we’re starting  to say maybe you are a syngapian and then i watch  

58:02your presentation and then and then there’s the  missense people who have been told that everyone’s   haploid sufficient but we’re not sure about them  so everyone’s got their own genetic report which  

58:12starts to form their identity within this syngap  community right and then i watch your presentation  

58:18and i kind of want to say well actually those  are all just highly highly educated guesses but  

58:24until we are able to run each of these mutations  through multiple assays we’re not exactly sure  

58:30and and i guess it’s a totally unfair  question but for the non-scientific person

58:37the the takeaway from your presentation is that  when when i describe syngap to people i’m like  

58:43it’s not a knife it’s a swiss army it’s a it’s the  biggest swiss army knife it’s one of the biggest   swiss army knives in the brain it does all these  things and we only know about a fraction of them  

58:50so honestly we’re still learning but when a parent  is thinking when are there going to be ASOs or  

58:55other genetic therapies available to to address  the lack of SynGAP in the brain and then they  

59:01watch this presentation from which they take away  actually every mutation is slightly different in a   number of ways sorry about my cuckoo clock where  do they go from here is it is it exciting that  

59:12we now have this level of specificity and we’re  going to know so much more about each mutation   or is it terrifying that everything we think we  know is actually just a a broad simplification

59:25yeah i don’t know i don’t want to  give false hope i as a biologist i  

59:33i we always expect to see complexity that’s not  surprising to us you know and we haven’t even  

59:38talked about things that sing up one has alternate  splicing isoforms and there’s a lot of different  

59:44things additions it’s complicated SynGAP on it  is not a simple protein as you say and and a lot  

59:52of proteins you know and the disease associated  are simplified down to a single mechanism and a   single term like haploinsufficiency is raised and  i don’t know what i i can’t as a the one during  

1:00:04the work and the experimenter that that’s not a  part of my world that’s we do the experiments we   see this complexity and that complexity doesn’t  surprise us at all however that complexity is the  

1:00:13starting point for us if we’re going to try to  to formulate a therapeutic that and decide well  

1:00:19either if we want to direct a therapeutic at  the molecular mechanism if haploinsufficiency  

1:00:24is the mechanism i can understand just just  raising it that’s that’s makes sense or i was   too much decreasing it but if it’s something  else i would try to target whatever that is  

1:00:35but then putting it into the context of a circuit  you might target the downstream circuit components  

1:00:42or you might find the pathophysiology and there  might be a way of targeting that pathophysiology   specifically or some other higher order symptom  so there’s many different ways of looking for  

1:00:52for therapy options but it requires a very  found understanding of what’s going on  

1:01:00so we’re at the you know there’s a lot of known  about singapore but yeah it’s it’s always a long  

1:01:08process yeah sorry my questions aren’t always   fair i’m just thinking out loud and one more  and then JR is going to use big words again but

1:01:20so the thinking that an asl that would just  upregulate both wild-type and mutant alleles is  

1:01:27is is likely to solve this problem at least for  the ptv’s but the more sophisticated genetic  

1:01:36therapies that are in the pipeline that might come  in and actually correct the mutation assuming we   can solve delivery or become more appealing in the  context of the complexity that your presentation  

1:01:46so beautifully illustrates right yeah i would say  that is very attractive and again that that then  

1:01:55highlights the ability to find those variants  that are likely to be causal from not that are  

1:02:01best candidates for that type of approach right  thank you thank you very much this is great jr  

1:02:10i was just going to ask if for some of  the protein truncating variants if you  

1:02:17well not even that for wild type if you would  uh look at some of the different isoforms

1:02:24yeah we’ve we’ve thought about doing that and we  we should so that’s on our list of you know sort  

1:02:30of stuff to do it just yeah it’d be really  interesting to see wild type and different   isoforms and see how they behaved against each  other in the same essays it might give some

1:02:42uh i don’t know it might it might just make  it more complex but yeah we have to think you  

1:02:48know should we take our a subset of of impactful  variants and put them in the different iso forms  

1:02:54see what happens and then move all those  towards a neuronal model might be a good idea

1:03:09okay we’re at the hour do we do we miss any  questions um my only other question that makes  

1:03:15me think when you mention the background um  some of our kids they will have the generic  

1:03:22report of sync app and then they have the buses  for others you know then it becomes more complex  

1:03:29yeah then and they may have a report that  is five years old then it comes to the same  

1:03:36thing you probably have to retest it at  some point see what is true bosses what is  

1:03:44yeah the background complexity  what is your thought about that  

1:03:49as an experimenter when we thought about uh  dealing with this you know we we always think   about doing ips cells as a model taking cells from  human individuals and and differentiating them  

1:04:01into neurons and using that as a basis for for  studying things and and that preserves the genetic  

1:04:07backgrounds and maybe looking at different genetic  the same mutation on different genetic backgrounds  

1:04:12um yeah it’s it’s that’d be an experimental way  the other is just a very large bioinformatic   way of um sort of trying to understand  how you know you know just statistically  

1:04:23how these different variants and different genes  interact well i think we’re just very far off  

1:04:29i mean i think this field we call functional  vario mix is at you know it’s just it’s really  

1:04:36early stages but in the future we’re going to  want to study every you know amino acid every  

1:04:43nucleotide in every gene and you know we’re  trying to move through autism associated genes   and eventually maybe we’ll have this huge data  set to say you know what’s the first way to say  

1:04:52what’s the impact on just that gene and then  the combinatorial impact is going to be very  

1:05:01tough and costly and time consuming to try to to  do that experimentally we’re thinking about doing  

1:05:08baby steps of doing just two genes but when she  started having combinations it’s just it’s huge

1:05:16can i ask a question and just how this research  came to be i mean PTEN is a big kid and then  

1:05:22SYNGAP1 and i saw safari so are you are we just  catching one chapter of a process throughout which  

1:05:27you’re gonna with safari support work through  multiple genes or is this the beginning of your  

1:05:33lab starting to become fascinated with SYNGAP1  i’m just trying to understand sort of where we are  

1:05:39in your relationship with SYNGAP1 yeah it’s both  those things i mean um so we do want to go through  

1:05:45many genes we’ve we started with PTEN and then we  did senegal but now we’re doing dirk 1a as well  

1:05:52and then there’s other genes that we’re moving to  where we now we’ve done covet over the last year  

1:06:00so uh but yeah every time we we get into a  gene we fall in love with it and understand the  

1:06:06complexities and every every protein is completely  different has its unique problems or interesting  

1:06:11complexities and i think once we establish a  platform we certainly are invested in developing  

1:06:19and using it and building on it yeah of course  we’re limited by by money and people time but  

1:06:27yeah it’s it’s it’s not something we just want to  go march through without stopping or it’s building

1:06:36but it is it is frustratingly fascinating  right there’s it’s so complicated and when  

1:06:42you add in the isoforms it’s there’s  a lot of interesting questions to ask  

1:06:47but you know and when i’m talking to hans and  jr they’re just like it’s endless they keep   throwing things at me i don’t understand but  my donors aren’t the people who support us are  

1:06:57off aren’t giving us money because they love  beautiful science although they appreciate it  

1:07:02they’re they’re asking the question  what what is getting us closer to um  

1:07:07therapy right and of course basic science informs  all of this and i’m and i guess i’m trying to  

1:07:15understand what the future looks like a little bit  right so in in three to five years when there are   genetic therapies will it be show as your genetic  report and you have a pathogenic mutation and  

1:07:26therefore you’re eligible or will it be let’s look  at your mutation let’s take a blood sample let’s   run it through 10 assays and let’s there’ll  be some score which will tell us a variance  

1:07:37amenability if that’s the right word  to a given genetic approach like what   in your opinion is you’re in this i think you  called it something variomics functional variomics  

1:07:48what is the future of giving genetic  therapies to young patients look like  

1:07:55yeah i raised this before which is i think if  the genetic therapies are safe and effective  

1:08:03you know given that you only want to use them  on the right targets and as we know it just  

1:08:11because you find a variant in somebody with with  a disorder it doesn’t mean that that’s causing it  

1:08:18because we at this point we know so little about  what those variants do even with a predictive  

1:08:24algorithm tells you something so it’d be very  important to have that pre-assessment say yes  

1:08:31that variant we think is very dysfunctional and  or we think this is the molecular mechanism maybe  

1:08:37you’d have a bioassay that can be performed on  somebody to show you that as well and then that  

1:08:44would be justify that as a target for genetic  therapy got it thank you anything else guys  

1:08:56thank you so much yeah and thank you for tackling  this huge amazing bit of biology it’s really it  

1:09:04really needs someone to to work hard on it and we  really appreciate that you and your whole lab are  

1:09:10going through and thinking hard  about all these things so thank you   well thanks for your your questions and thanks for  your initial question by email that really allowed  

1:09:20me to switch my focus for my talk fantastic  talk yeah this was amazing yeah very very  

1:09:29yeah i can’t wait to rewatch it and really look at  the graphs more thank you thank you again thanks