 Welcome back in Halle with ChaosZoneTV. The next talk will have interactive elements, so here are the hashtags again. We're on Mastodon with and Twitter with the hashtag RC3ChaosZone and on the IRC channel in Hackint which is RC3-ChaosZone. Alright, Lisette will now speak to us with a talk called What the Health Beyond Genome Sequencing. Since the 80s the human genome projects had goals to technical and ethical goals to understand the human genome. In recent years these goals have been achieved and humanity got profit immensely from what the sciences and the technology, the methods could be developed through the project. Lisette works at the bleeding edge of what it is now, a hard data science. We're very excited to hear about the considerations and the practicalities of advancing the biology even further. Okay, thank you very much for the introduction. It's my pleasure to give some insights into what I've learned throughout my studies and what I'm now actually also working on. So thank you for providing me the slot. I was a little bit surprised when I thought, okay, now I actually have to give the talk. So please forgive me if I'm sort of nervous, but stay with me and thank you everyone for watching and for filling in the survey beforehand. And you will have another option to participate in the poll later on. So I have some things to announce first, which would be about the content. So it will all be very abstract. So we are talking more about concepts than about actual disease and suffering. So there will be no photos. But yeah, the general theme is about medical examination, everything clinical about the patient assessing somebody's disease and disease risk. And also going into the more severe conditions of which you might die. And we also touch upon family relationships. So just so you know, yeah, it will come back every now and then. So just for everyone to be aware. And then also, yeah, I need to disclose that I'm an employee of a company that does work on marketing. Genetic tests. So that set aside, this is not, this is not any kind of advertising talk. It's really about what is actually happening technology wise. So I want to give you the insights into a little bit of the technology, how it came about and where we are now. And also try and give you an overview of what are the options in terms of genetic testing for various utilities and raise awareness just for also the ethical issues that might arise from what we can learn from our DNA. So this is enough of the prologue. Let's go right into looking at a patient, which is classically done from the outside. So we want to know what is different about this person or patient. And yeah, there are really layers of information. And you always assume that there's a relationship with a condition. So be it a rash that you see on the outside or as welling that the doctor can feel or something that they learned from interrogating the patient. And then there's a bit of a borderline outside, inside test, which would be bodily fluids. So if you test urine, saliva, blood, you already look on the inside. So what's happening inside of the patient, the metabolome, what's going on in terms of small molecules that you might detect with one or the other test. And also what you can see on the inside is broken bones or fists that shouldn't be there. So for that, we use imaging where x-ray is the oldest and then there's magnetic resonance and PET scanning. So these are like the cool advanced additional layers where you can look inside of the patient. And then of course, there's DNA. So if we look even deeper and inside each cell, you will have the genetic code of this person. So to tell how they are different on a very small scale. So that is the dogma of molecular biology that you go from DNA, which is your genetic blueprint and then certain parts are transcribed. So the cell makes copies of the DNA, which are then called RNA just because it's a different chemistry. And these are then translated into chains of amino acids. So there's a code, which amino acid should be attached to which one and then you fold it properly and then you have a functional protein. And then now why you sequence the DNA is because you assume that there's a mistake made, which then leads to a faulty protein. And then in the end, something in your body doesn't work. So, yeah, it's a very simple concept, if you will. And then, yeah, what we check in the DNA and in the RNA is about 20,000 protein coding genes. And then there's also different types of RNA that do not code for proteins, but that regulate other stuff. So that the correct genes are actually transcribed and translated. So that's an additional 20,000 to 30,000 potentially more. And so if you combine any of these to see like a certain signature of a person, you already have billions of combinations. So as you can imagine, there are many, many, many signatures possible. But yeah, which of these will actually tell you something about the patient. So let's go back to how we sequence the DNA. So it is actually very simple. All of our usually 46 chromosomes, so the 23 pairs are made of a double-stranded code, which is the DNA. And then you see here in the unfolded region that where a gene is starting, it usually starts with ATG. And these are caramidin bases. So you have here in the chemical little inset the A and the T, which form a pair. So the red thingies in between are hydrogen bonds that keep them together. And A and T always want to be together. And C and G always want to be together. C and G actually form three of those bonds, so they're a little bit more stable. And so as you can see, this double-stranded DNA is hence always inverted on the other strand. So we call it the complementary strand. So if you have ATG on one strand, you always have TAC on the other. So you only sequence one. And we defined the direction of the gene because we know in which direction it makes sense. Because in only in one direction, you can then make a protein out of this code. So enough for the chemistry and the principle. So we really want to know and to map where on each chromosome which letter occurs. So you can imagine that this is quite an adventure and takes a lot of effort. And actually it has also started very early on in the 70s. So maybe you have heard of Sanger sequencing. So that was the first generation of sequencing from 1977, where you essentially cut the strand in little pieces. And you know which one ends with an A, ends with a T. So you have all kinds of fragments with different lengths, which run over a gel, which is not that important. But it is also called capillary sequencing, which then helped finding the first human disease gene, which is called the hunting team. You might have heard of the disease where it belongs to Korea, Huntington. And so this was the first association that was really confirmed that, okay, you have a defect in a certain gene, which directly translates into a disease phenotype. But this is very rare. So usually it is a lot more complex and we will also get to that. So the capillary sequencing still lasted for a while. So 10 years later you had really cool instruments for the first time from applied biosystems, so that you can sequence a little bit quicker, but still far from looking at the whole genome. So that was then planned starting in 1988. They defined the goals for the human genome project, which would then take from 1990 until 2003 to complete one full human genome. So full in the sense that it still had gaps, so there are some regions which are tricky to sequence, so these gaps were filled later on. But still, yeah, this was a huge undertaking which cost about two to three billion US dollars. And eventually in 2000 they announced that they had first dropped of the human genome. And then it got published in 2001 in the two big scientific journals, Nature and Science, both on the cover, the human genome. So that was and is a big step. So yeah, that's just crucial to know what we are looking at to have a map of our complete genome, where then you can map other people's sequences to as well. So that's what started also in 2005, but then for different types of cancer. It's called TCGA from the cancer genome atlas. And it also lasted for a couple of years, but then they were much quicker in sequencing because 2005 was also the year of next generation sequencing machines. So nowadays, we don't do Sanger sequencing anymore, or rarely. We usually rely on heavy high throughput parallel sequencing so that you can sequence a lot more different pieces, so to say, at the same time and with very high accuracy. So essentially, this means that we now have access to 3.1 billion base pairs, which were first collected during this human genome project. And this nice advertisement when they were looking for volunteers is really cute, actually, because they also say here that the outcome of the project will have tremendous impact on future progress of medical science and lead to improved diagnosis and treatment of hereditary diseases. Volunteers will receive information about the project and sign a consent form. No personal information will be maintained or transferred, and a small monetary investment will be provided. So, yeah, they were promised that their data would be kept anonymously and also they collected blood from female volunteers or sperm from male volunteers. And then they collected a lot more samples than what they would need so that in the end, you couldn't tell anymore from whom the genome was actually derived. And there was one volunteer at Rosewell Park and hence called RP-11, who happened to have exceptional quality sequencing reads. And then so the first human genome was mainly based on this one person. And we have multiple new versions published of the human reference genome today. It's version 38 and still about 70% are untouched from this first genome assembly. And a small thing about the cost. So I mentioned that this was a really costly project, two to three billion dollars. And now we have actually cracked the $1,000 threshold. So it is possible to sequence a full human genome for about a thousand bucks, which is remarkable. So this is really an enormous drop in the cost just because the technology made such a big leap when we came to the next generation sequencing. And also, one genome, if you have it sufficiently covered so that you are sure about which base pair is at which position, then you have about 180 gigabytes of raw reads. And if you align them to the reference genome, which is of course now your atlas, if you will, so you can put all the reads to the correct place. And then this is called an alignment file, which is about 80 gigabytes. And if you then only keep the positions where something is different from the reference genome and you compress it, you are left with about 5% of that. So four gigabytes per person, storable, nice little genome. OK, so this takes me to the first poll, which is on simple vote. And a couple of people already have participated in the monkey survey. Then, yeah, you don't have to do it again now. But the vote link will also be in the chat. And you also just fill in any name combination of letters, click OK. And then you can answer the first question, which I present here. So this is just three statements about sequencing a full human genome, whether you believe that it has replaced fingerprinting in forensic investigations, whether you think that it gives you all the clinically relevant information for any patient, and whether you think that it is cheaper than a full body MRI scan. So yeah, we will get to the results in a bit. I will just continue with a couple more slides and then we can see what you guys think. And I'm really curious to actually hear that and see it for myself. Let's see. So if you think in terms of complexity, we have already touched upon Korea Huntington, which is a single gene, essentially, that gives you a full blown disease if it's not encoded properly. And then you could think of other diseases that are encoded by a couple of genes, where you can think of breast cancer, where a couple of mutated genes can give you a much higher risk than average population. And also in Alzheimer's disease, we see that hereditary component, brought about by a couple of genes again. And then more general in terms of unknown diseases, you can ask gene panels or full genome sequencing to help out. And it gets more and more fuzzy, but more and more also tests are available if you want to go to a prognosis for this or that condition or to the correct treatment choice. So I'll try and give you a couple of more examples. But only after we have talked about the cancer genome atlas TCGA. So here, that's also a lot of data. So they claim 2.5 petabytes were collected in the phase it was running from 2006 to 2014. And yeah, in total, 33 different tumor types. And they did not only look at the DNA and all the mutations, but also RNA and also proteins. And also different info on the patient's survival and treatment data. So that is a huge pool and resource of data where people are looking at and finding signatures of patients with less or more advanced cancers with patients that progress through treatment or not. But it's all, yeah, you still really need to take it with a pinch of salt. Because, for example, since 2006, treatment of cancer has changed tremendously. And you cannot just use any signature that you took from the data from TCGA and extrapolate for today's cancer patients. So that's a bit tricky. TCGA is still vastly used. But then, yeah, I would propose that you should rather use it for validation. So you find something in current data from today's patients. And then you can check whether this was also seen in the TCGA data and not the other way around. But let's get to the results of the poll. Let's see. Can we go there? What happens? Oh, nice. What's the score? 7.3. So you mostly agree that full body MRI is more expensive than the full genome sequencing, which is true. So, like I said, the full genome is now about one thousand dollars, also one thousand euros. And the full body scan in the MRI will cost about two to six thousand euros roughly. And then this one with the fingerprints I have made up. So, sorry to fool you. This is not done yet. And it also cannot potentially give you all clinical relevant information about the patient. So, nice. Thank you for participating. And also, I've checked the survey monkey. And also there I have managed to fool some people into believing that it's possible to replace fingerprinting with full genome sequencing. But that's not true. Sorry. So let's go to another level. So not only the DNA sequencing is interesting, so then you have the map. And on the protein, sorry, on the DNA strand, you know, for example, where there's a different letter, if you will, then in the reference genome. And then this mutation might be in one of the regions where the DNA has stored the code for a certain protein, like protein one or protein two. So the code might be different, but also it might be different how many copies are made. So this is an example here where gene one and two are equally often transcribed. And then there's these transcripts, which we call messenger RNA, about equal amounts. And this is, let's say, the state how it should be in a healthy adult. And if you think about any condition like a cancer tumor, then it might get deregulated. And the cancer, for example, then does this and only makes very few copies of gene one and a lot of copies of gene two, which might lead to effects like bigger growth, faster growth, bigger spread into the tissue, which would normally confine the tumor. So that is also one level of regulation. And that you cannot usually capture with DNA sequencing or whole genome sequencing. For that, you need to check for the expression, which you do on this level, on the RNA level. And then you have, they call it differential expression, which gives you this kind of picture after the analysis. So you have samples, vertical and then horizontal are the genes. And you see that if you compare the samples, some genes are more expressed, which is red and some genes are down compared to the others, which is green. And then you can find clusters of genes, a group of genes here in the red bar, where group one, in that case, a certain kind of breast cancer is highly upregulated. And most of the people that belong to group two, different kinds of breast cancer, have lower expression of that gene. And in the blue cluster, it's the other way around. So that gives you an idea of, OK, you can maybe use one of these genes to differentiate between the two groups. And if that helps you to determine what treatment they should get, that's of course super useful. And then you have something like a genetic biomarker. If you have multiple genes, then you usually call it a signature. And so these genetic signature tests can tell you, are you at risk of a certain disease? They can help diagnose or get to the exact subtype of your disease. They can help use the correct treatment or monitor whether the disease actually responds to the treatment, whether anything changes back to normal. And also it can sometimes be useful to give a prognosis for the disease progression. So in the end, you always need to wonder what is the added value of such kind of testing on top of the clinical variables that are already existing. And does it give you something actionable? Can you do something with the knowledge that you gained from this testing? So there we are already at the problems with genetic testing. So that would be the second question that you can answer again on simple vote, please feel invited to help me understand what you think. And here it's just for you personally, the question whether you would want to know whether you are at risk of a genetic disease. And would you want to know if you had to pay for it and then slide it to the right if you're willing to pay or slide it to the left if you're totally not willing to? And then the second slider is the same question, would you want to know if you got the results for free? And then to the right is yes and more to the left is no, absolutely not. So again, I will just move on and you can take your time answering that one. So to give you a bit of a feeling for what is at stake if you get into testing for genetic risks, it's of course good to know your family history of disease. And also, if you're planning to have children, for example, would you want them to know that they potentially carry a certain risk or not? Then health or life insurance might have an interest in knowing what people's risks are, what they have to expect. So there are certain instances where they are eligible to know and certain instances where at this moment in time they absolutely are not. So this is something that's probably going to change in the future the more we know, the more we want to use that knowledge. And then there's the problem that some genes are very often found to be up and down regulated and there seems to be a difference, but it's just, yeah, in the nature of those genes. And we have sometimes multiple signatures for the same problem and then doctors and patients just don't know what to choose from. So I'll go through some of those issues in more detail. I have mentioned TCGA before and this cancer genome atlas is really a limited source that is now exhausted, but it's still oftentimes used as the silver bullet. So let's see if we already have votes. So that would be the second one. Okay. So if you could know your genetic risk and you would get it for free, then most people are inclined to say, yes, I would like that very much. And if they had to pay for it, then it seems to go more towards no, but it's actually kind of neutral, which is surprising. Yeah, I would have thought that you would all say no, I don't want to know, but that was just my assumption and I was apparently wrong. Cool. Thank you. Paul number three, third question is about a commercially available DNA test, which is not actually sequencing, but they use like a panel of mutations that are now known because we have already sequenced thousands and nearing a million complete full genomes. And yeah, I was wondering whether you would know. So that's question number three, what institutions they partner up with. So this DNA test is called 23andMe. And if you don't know what it is, then there's also an answer option for this one. No clue what it is, what it does. And for the rest, yeah, I propose that they work together with Broad Institute, that they work together with GlaxoSmithKline, GSK, and they got 300 million US dollars from them. That they work together with general practitioners in the US that they got subsidy from Google, four million US dollar or and Amazon, nine million US dollars. So, okay, let's see what you think or how many of you don't know the test. And in the meantime, I'll present two cases to you. Where genetic testing would play a role, like for instance, in the case of an healthy adult, where the dad was diagnosed with this heart condition, hypertrophic cardiomyopathy, where the heart tissue gets scars and at some point it cannot pump properly anymore. And so if you have one parent with that disease, you have a 50% risk that you have inherited those difficult genes from your parents. So this healthy adult and their siblings got the offer to get tested. And so the costs are covered by the health insurance, but there's no cure for this condition. So you can have a stricter surveillance and you can get access to early treatment if you develop symptoms, but other than that, yeah, it's still just a risk gene, so to say. So if you know you have the gene, it doesn't mean you will get the disease, it just means you have an elevated risk. So it's really hard to grasp. And this is one case where at least in the Netherlands, the life insurance would be eligible to know if you got tested and you do have that gene. So in the end, this person said, no, no test, please, I will just go see a cardiologist every now and then, have it checked nonetheless, but I don't want to know if I have those genes. Okay. Second case. Yeah. So that's an infant delayed in development. It was still a bit fuzzy, like what should an infant be able to do or not do at the age of one. But then the parents started observing seizures in that case. It was absences, so it was not cramping, but just, yeah, very absent. So eventually they got access to tests, genetic tests where distinct genes were analyzed, nothing was found, then panels of genes with increasing size and nothing was found, and then the whole genome sequencing was done, and then you always have to compare to the parents and essentially parents and child were tested and the child had a mutation in the gene where the parents had nothing. And yeah, it was just a very rare, excellent mutation. And eventually they now know what is going on, which was only due to the possibility of whole genome sequencing. And in the end, the parents also said, yes, I want to know what else is found in this whole genome sequencing. So that is actually case three, where one of the parents is the carrier of a mutation in a protein that when it's faulty, when you get a faulty version from both parents, then you will develop this condition cystic fibrosis. So that is really good to know when you are a carrier of this, and also your future kids can get tested to see whether they got this faulty version from you. So let's have a look at poll number three. This is here. So the DNA test, 23 and me, let's see, where's the, I have no clue what this test is. So this is just a four. Okay, so not that many people voted for this one. 29 votes. Oh, well, actually 29 votes. And then what you thought it would do, so you'll have here, you approve of it working in conjunction with general practitioners in the US, which is not true. Sorry, yes, it did get subsidy from Google for a million US dollars in the very beginning. No, no, no, um, saying our sequence saying, um, yes, GSK 300 million, they want to use their data to find new drug targets. And I also made this one up so Amazon did not give any money to 23 and me, but you can order through Amazon. So that's possible. Okay, thank you. And I'll think I will wrap up after just presenting this problem here quickly. So breast cancer is one of the pioneering fields of genetic testing. So you have five commercially available tests that can tell you what type you have, what treatment options would be best for you and what your prognosis is. So you really need a well informed team of doctors if you want to make use of this. Okay, um, I'll skip a few slides. I mean, validation is important, takes a lot of time. And I think in the future, it's not only going to be whole genome sequencing, but there will be a lot more to it, like the immune system and your gut microbiome and everything, which is in there is also, of course, influenced by outside factors, what you eat, how much sunlight. You get how much you move. So this is also, um, already available, this data from your smartwatch, for example. So I think in the end, if we get to personalized medicine, this will also play a role and to recap, um, if you sequence the whole genome, this is not the same as ordering any tests online where you also might run into data security issues. With tests like 23 and me, and that's also not the same as a disease signature. And then, um, yeah, if you have a new, cool diagnostic signature that is published, it might still take a long time and couple of validation studies before it actually enters the everyday clinic and you get it reimbursed from your health insurance. And for this, it also needs very well trained physicians and informed patient and family. I think there's no way in stopping this, um, but that's just my take. So we will see a lot more from the molecular side of things in the future. And, um, these are also to be retrieved online. So everything, all the tests that are registered also, you can filter for countries for Germany, for example, and then you see even which university clinic offers, which kind of testing. And, um, if you ever hear the term liquid biopsy, that's usually a blood sample where, yeah, all kinds of things are measured. So you have DNA in there, but you also have metabolites in there. You can have little fragments of cancer cells and cancer derived DNA. So this is something that's coming forward more and more, that you just need a blood draw and then, yeah, you have a lot of insight, not only the whole genome, but even more RNA sequencing data, for example. So thank you very much for inviting me for listening and I'm happy to take your questions now. Once again, the social media hashtags on Mastodon and Twitter are RC3 chaos zone without a dash and then on IRC, on Hackint, the channel is RC3 with a dash chaos zone. Do we already have any specific questions? Anything people would like to know? And a targeted gene modification with CRISPR and Cas9 is not even allowed on plants and animals in the EU. Do you think there will ever be gene therapy for humans? There was gene therapy. So, for example, I'm not sure whether it was a type of leukemia or an immune defect where they tried to cure children with gene therapy. So there were clinical trials, but something went horribly wrong. And I think actually one of the children suffered so much from how they inserted the gene that it developed a type of cancer. But I'm still hesitant to say that this is the end of gene therapy. So it has potential in various severe cases where there's no other option. But yes, it's also true that we don't really know what we're doing at the moment. So there's a lot more research needed to make sure that there's no off target effects if you cut out a gene and put in a new sequence. So, yeah, no, I don't think we can guarantee that as of yet, but it's not unthinkable. All right. Sounds like the technology isn't there yet for a couple of years or decades. Oh, well, I think the technology is there. It's just not secure enough. All right, I see. So yeah, it's done in the lab big time. But yeah, then we don't usually use humans, only a cell line or something that is easy to control. All right. And then, uh, genomic methods for tests, for example, for diseases such as COVID are targeted tests, for example, the PCR test. Do you think the testing for infections might shift to be more exploratory approaches? For example, through sequencing instead of target PCR? Yeah, that depends. If you have a suspicion that the infection has reached the bloodstream and you're close to sepsis, then it might be your last resort to make a whole, uh, yeah, just sequence everything that is in the blood. But then you need to be, of course, aware that the majority will be human. So you need to filter out a lot. And then what is left, you might be able to map to certain microbe genomes, which are also pretty well annotated. So I'm not sure about nasal swabs or something like that, where you can find out which flu you have received. So that doesn't really make too much sense to me unless you have good treatment options. But for example, tuberculosis is one disease where you do sequence the germs now more and more, because a lot of strains of these bacteria have multiple antibiotic resistances. And then if you start treating with the wrong antibiotics, you are really screwed. So there, yeah, it's already well established that the university clinics at least sequence the strains before the patient gets treatment. Interesting. Yeah, sounds, sounds very cool. All right. Thank you so much, Lisette. Very inspiring. You're welcome. It was a pleasure. I hope I could convey the message, just be aware of your genes and your data. So yeah, that's just there's a lot of potential in there. But yeah, of course, we couldn't be, we should not be careless. So yeah, that's all from my side. Thank you. Thank you so much.