 Good afternoon, everyone. It's my pleasure to present some of the findings from our Insight project at UCSF called NB Seek We have no conflicts to disclose and I also want to say what I'm reporting is Results that a whole wonderful team has put together and everything they have contributed it Contributed is correct. And so if there are any mistakes, they are my own So NB Seek was formed to address these questions as you just heard that were posed By NIH first could exome sequencing replace the current mass spectrometry That is performed by public health laboratories as a newborn screen for inborn errors of metabolism and second Maybe it couldn't replace but could sequencing augment the information that is obtained Through current newborn screening programs and actually improve case resolution and public health outcomes So newborn screening has to work as a public health program It focuses on conditions which are urgent their infant onset and Treatable disorders that are not detected except by newborn screening The demonstrable public health benefit and practical considerations have allowed newborn screening to be performed Without explicit parental consent Medical need rather than technology should be the driver of any Population-based screening including newborn screening. So those are our starting principles also Just to describe the disorders that we chose to look at which is the inborn errors of metabolism in California These are rare genetic defects that cause serious disease and infancy, but that are treated And successfully treated with special diets and other Treatment modalities the list includes 48 inborn errors of fat metabolism organic acid metabolism and amino acid metabolism and about 150 cases of Infants with these disorders are identified each year in California which has the largest number of births of any state in fact It's a one-eighth of all the babies born in the US are born in California And newborn screening by this MSMS or mass spectroscopy is is extremely effective you can see here that The sensitivity is overall about over 99.3 percent and the specificity is also well over 99 percent the sensitivity is important because newborn screening really should not be missing infants and and it's Really tragic if an infant is missed and the specificity is very important because there's a lot of anxiety Involved in telling a family that their infant may have one of these conditions And there's a lot of medical cost in terms of specialist effort and testing that goes on so you don't want to have a lot of Infants and their families involved in this if they don't actually have a disorder So what are the resources that we brought to bear? Well, California, I told you already has a lot of infants born every year and it's also an extremely diverse population so that is Quite interesting for example in cystic fibrosis It was found in California that the typical panels that were good for Northern European Caucasians Just didn't work in California where so many people have different ancestral backgrounds We were able to take advantage of California's archived dried blood spots and these are the Blood spot on a piece of filter paper that is used for the metabolic screening and other newborn screening tests And in California the leftovers from this testing are stored in a biobank For potential future use and I'll get into that in detail because it's very important Follow-up data from the metabolic specialty centers is also Tracked in California and put into a central electronic database So so we have follow-up data as well as newborn dried blood spots And we have a multidisciplinary team with expertise in genomic sequencing and analysis genetics Ethics and clinical diagnosis and management of these inborn errors So the California Biobank is an Invaluable and irreplaceable resource the dried blood spot residuals have been stored there since 1982 and The samples can be made available for research to improve the health of women and children and What's listed here is the website of the California Biobank where the Regulations and and law is establishing the bank and its use are described in detail The purpose of this Biobank is to foster development of new and improved newborn screening tests and other modalities for the public good Projects using the Biobank materials have to be IRB approved and the samples have to be de-identified to protect the privacy of individuals of the babies who are born and Good data stewardship is absolutely required and monitored with great care to protect the individuals Privacy and at the end of the study Remaining samples and large DNA data files must be destroyed Any individual may request to have his or her sample removed and destroyed and such requests are all honored with documentation so we thought a lot about the distinction of different contexts for genetic information and as everybody is aware there's been a lot of There's been a big increase in the amount of Genomic Sequencing that is being done And it's being in some cases offered directly to the public without a doctor in between but clearly The context has has been different in different situations and we mostly think about the diagnostic Content context of an infant who has some abnormal phenotype and we don't know an underlying cause so we look for Genomics and genomic sequencing for an answer so this is an individual patient who is Being evaluated by a doctor with an abnormal phenotype that is the start and the driver of the analysis But public health newborn screening is completely different in this context. There is limited or no phenotype Almost all the infants are unaffected and those who are affected actually appear healthy because they're pre-symptomatic So the enthusiasm for success in the diagnostic context Shouldn't be allowed to carry over willy-nilly into the newborn screening context They're really quite different. I think many people are already familiar with whole exome sequencing and and it was Described in the introduction to this session so in this technology DNA fragments are captured and sequenced from the coding portions of the Roughly 20,000 human genes and this represents less than two percent of the entire genome and excludes introns and The intervening DNA sequences and some of these we know are actually important For regulating genes, but we don't know much about how they're important So the exome is sort of a shortcut. It's not expected to be completely Effective, but it is easy to come by and and useful, so we this is the Technology that we settled on and the newborn the NBC group Elected also within this exome to analyze only 78 genes that are known to be involved in inherited metabolic Disorders on our list and we call this an exome slice So this is actually the largest to date whole exome sequence study of an unbiased Cohort of patients with inborn errors of metabolism and so we think that this study can Establish a benchmark for the capabilities of exome sequencing in the newborn screening context these are just some numbers of Samples in our study that I want to go over very briefly. We we first requested about 1700 samples and We included those with a known Diagnosed inborn error of metabolism including nine who had been missed by the newborn screening Process, but were picked up Clinically later in life. We also included Nearly 400 false positives and then when that After the request was made we had to eliminate 538 of them for various reasons we ran out of money and also some of the samples Even though we attempted a sequence the exome didn't pass our quality control metrics so we ended up with nearly 1200 samples to analyze and of these 805 did actually have a proven inborn error and there were 385 false positives by the mass spec test and We divided the samples into an initial small validation set and A and then the rest of them were our test set and the validation set was used by Dr. Brenner and his colleagues at Berkeley to develop a Screening pipeline for analysis and and this was different from the typical diagnostic Pipeline and and we really don't have time to go into it unfortunately, but These are results that came out at the end of the day from the test set of individuals who were actually affected with a disorder and You can see that the exomes did identify all kinds of Known and and new predicted mutations. You can see there some of them were actually already listed in ClinVar or HGMD databases others were Very rare variants some were predicted to be damaging by introducing Stop codons and so on and then at the bottom you can see So of those 674 we got 571 right, so I would give us maybe a B plus for that because we missed 103 and I want to just give you a few examples of What we did right and what we missed So here's one we got right just to show this is the PAH gene which is mutated in phenyl ketoneria and You can see in this list of variants that our exome found a lot of variants in this particular Individual and that's quite typical the genome contains a lot of variation, but those Two top ones shaded in pink are actually known pathogenic variants and so When our pipeline identified two Clearly pathogenic variants it flagged the sample as positive and so this sample was correctly diagnosed Here's one though where we didn't do quite so well and so this was a case of Gluteric aciduria type one and the gene Had a lot of variants and you can see the variants listed there, but They're either very common in which case we thought oh that's ridiculous can't be or we'd have people walking around with GA type one disease all over the place which we know we don't have and so we really rejected these variants based on Their frequency or some of them were actually known to be benign and we didn't come up with anything that Could explain this patient's phenotype But the patient did clearly have this disease so we missed this one Here's another one another instance where our exome wasn't Adequate and and missed two cases here of isovaleric acidemia and We wonder why that was and actually took some of these samples a subset of our of our exome cases and Submitted them to whole genome sequencing and you can see in these two instances. There were genomic deletions and exome sequencing is very insensitive for picking up this kind of defect You can see those sort of clear areas of white spots indicate areas of poor coverage and the first Individual had a homozygous Deletion and and you can see on the top. I don't know quite how this pointer works. Oh So here's a diagram of the gene and this is the deletion in this Individual and you can see that the This section of the gene that the five prime end of it was deleted And in the second individual there are actually two Deletions a little one here So there's no coverage in the exon 12 area and a larger deletion representing this area Low coverage across the gene So so insertions and deletions Exomes are not great for Here's another one we missed and and in this one we actually had we were good in in one case we got a Variant that was flagged that is known to be Well, let's see clean bar thought it was pathogenic and and htmd at least questionably a disease mutation and This Okay, so and this is the the The diseases MCAD here that we're looking at and we required our pipeline to find two variants in order to flag an individual because otherwise we'd end up with 30,000 or so Cases a year, which is just not reasonable and so We didn't find another variant that we could Identify so this one was not called out But in it turned out that the person actually had disease and when we went back and examined the sequence we noted that there was a Variant here just upstream from this exon and It was a potential splice mutation because of the Lariat formation and the branch point a but this the all of the prediction programs gave the branch point a role to different a Residues in this case and so we had to actually resort to a research lab That made a construct to check whether splicing Did or did not happen when this variant was introduced and Sure enough it turned out this variant is critical for splicing Slicing therefore would not occur in this allele and that did explain the patient's disease even though the screening Couldn't have picked it up So just to summarize this our overall exome sensitivity was 88 percent across all of the diseases, but but What's important here is that some of the genes did a lot better than others. So here I'm Indicating in green the ones that we got right and in brown The ones that we missed and you can see that even for a common Disease like PKU or very common like MCAT We we were unhappy that these cases that had been picked up by Tandem as spectroscopy were being missed by exomes So this leads us to think that Sequencing exomes alone would not be a good replacement for Mass spec newborn screening So what I can conclude here is that the whole exome sequencing worked as well as the mass spec screening for certain disorders and Actually in some cases even better because some of those false negatives were identified by our exome pipeline But the Exomes may of sorry the exomes may also work well in some cases where there is no Mass spec screen so we have to think about other genes We would like to screen for and don't have a test right now and maybe sequencing can be such a test however for many conditions the exome sequencing was less good and We really do have to consider gene by gene and This is leading to our final conclusion that sequencing could be a helpful second tier test Following a positive newborn screen to help reduce False positives in some of the conditions that have a lot of false positives now So what have we learned well one thing is that genetic tests do not guarantee That one can identify a disease perfectly and in fact a third of all the variants we found had never been Seen before and we're not put in any of the databases and that just reflects I think California's diversity that our our collection of variants worldwide is lagging and California is an indicator of that The exome analysis was insufficient to identify pathogenic variations in populations with diverse ethnic backgrounds and Also, we found that the notion of a variant being pathogenic is actually an oversimplification For these autosomal recessive disorders because it turned out that sometimes Diplotypes really Had a big influence on the final phenotype and and what I mean by that is that Sometimes a particular Variant on one allele with a particular one on the other allele Turned out worse than you would have expected For example worse than somebody who is homozygous for either one alone or sometimes Better than expected and so it really opens a new can of worms that we have to not only examine things gene by gene and variant by variant but by combinations of variants with each other Okay, so Overall we say at this time whole exome sequencing alone is unsuitable as a soul newborn screening modality for inborn errors of metabolism for selected disorders. However, it was just as good as msms and The sequence information from our whole exomes could reduce false positives in the tests already being done and could facilitate an Accurate and timely case resolution So this is our NBC team. I'm very proud to be a member of them and That'll conclude my presentation Thanks, Jennifer and Just a reminder for those who are listening in if you have any questions, please send them into the email address That's available on genome TV And with that I will turn it over to Cynthia Powell who's going to present from the UNC Chapel Hill site You