We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Deciphering the mysteries of human genomes

00:00

Formal Metadata

Title
Deciphering the mysteries of human genomes
Title of Series
Number of Parts
131
Author
License
CC Attribution - NonCommercial - ShareAlike 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
Have you ever wondered why we are the way we are? Why some individuals develop diseases while others remain healthy? And what does Python have to do with all of this? Join this talk in which we will explore the interface between biology, technology and medicine, in the context of the research of rare genetic diseases. Learn what the Moore’s law has to do with advances in genetics and medicine, or why bigger is not always better.
Magneto-optical driveRootCNNPrinciple of relativityThermal expansionStatisticsComputerVideo gameInteractive televisionCartesian coordinate systemNatural languageMathematical analysisNeuroinformatikAreaComputer animationLecture/ConferenceMeeting/Interview
Cellular automatonNatural languageSequenceCellular automatonMereologyMilitary baseEntire functionHelixSelf-organizationAtomic nucleusSet (mathematics)Natural languageData structureACIDInformationComputer animation
Service (economics)System of linear equationsLibrary (computing)Point (geometry)Graph coloringComputer animation
Design of experimentsPhysicsHypercubeOrganic computingCodeTranslation (relic)Cone penetration testTrigonometric functionsCloud computingSheaf (mathematics)Scale (map)ProteinCellular automatonVirtual machineProteinInformationMereologySpeciesWebsiteRegulator geneFunctional (mathematics)Atomic nucleusExecution unitTable (information)MathematicsData structureGraph coloringDifferent (Kate Ryan album)String (computer science)CodeWater vaporComplex (psychology)Membrane keyboardSelf-organizationScaling (geometry)Computer animation
OvalLetterpress printingCellular automatonSystem of linear equationsExecutive information systemType theoryContent (media)Cellular automatonBlock (periodic table)BuildingVideoconferencingVirtual machineAtomic nucleusProteinData structureMusical ensembleMereologyDirection (geometry)Group actionComputer animationPanel painting
Router (computing)CodeError messageSpacetimeRead-only memoryComputer programmingBounded variationSoftware bugError messageDifferent (Kate Ryan album)Computer animation
Service (economics)Organic computingPhysicsHypercubeCorrelation and dependenceSequenceProteinACIDSubstitute goodCodeShift operatorInsertion lossFrequency2 (number)FamilyMereologyCausalityObservational studyOnline helpExtension (kinesiology)Data modelCellular automatonFunction (mathematics)PseudodifferentialoperatorCodeMathematicsFunctional (mathematics)Beat (acoustics)Sound effectFrame problemReading (process)Radical (chemistry)DivisorComplete metric spaceMereology1 (number)ProteinSelf-organizationNatural languageInsertion lossCausalityContext awarenessRevision controlFamilyFrequencyCellular automatonEndliche ModelltheorieVideo gamePlanningMultiplication signHypermediaArtificial lifeCASE <Informatik>Bounded variationModal logicInformationEvoluteComputer animation
Moore's lawMathematicsSequenceReading (process)Projective planeElectric generatorMultiplication signPhysical lawComputer animation
Router (computing)Arc (geometry)Coma BerenicesOrdinary differential equationMagneto-optical driveSurface of revolutionNatural languageSequence2 (number)Different (Kate Ryan album)Electric generatorVirtual machineMathematicsVideoconferencingNumberProteinLengthPersonal identification numberArithmetic meanOcean currentReal-time operating systemDuality (mathematics)Sampling (statistics)Computer fileInsertion lossNeuroinformatikComputer animation
SequenceSpacetimeComputer animation
System callMathematical analysisSystem identificationSample (statistics)Computer fileLogic gateSequenceFile formatError messageTexture mappingWikiInformationField (computer science)Data typeBinary fileSheaf (mathematics)Template (C++)Query languageEmailFlagInsertion lossPosition operatorElectric currentLine (geometry)First-person shooterDigital filterDatabaseContext awarenessRankingPredictionFunctional (mathematics)Focus (optics)Bounded variationOntologyPoint cloudSupercomputerProcess (computing)Design of experimentsNetwork-attached storageRecurrence relationEndliche ModelltheorieSequenceError messageMappingInformationPosition operatorInsertion lossSystem callFamilyMultiplication signCellular automatonBasis <Mathematik>Raw image formatDatabaseLibrary (computing)Data miningCartesian coordinate systemReading (process)Film editingCuboidDifferent (Kate Ryan album)Form (programming)Computer fileResultantElectronic mailing listBounded variationVirtual machine1 (number)Observational studyEmailValidity (statistics)Mathematical analysisLatent heatBefehlsprozessorLevel (video gaming)Web browserSampling (statistics)ReliefFunctional (mathematics)File formatLaptopNeuroinformatikFrequencyMilitary baseComputer animation
GradientOnline helpPlanningFamilyDatabaseLogic gateSoftwarePhysical systemData managementActive contour modelOpen setNatural languageRouter (computing)Core dumpMathematical analysisMeta elementDifferential (mechanical device)Computer-aided designFreewareThomas BayesObject (grammar)Row (database)WikiFunction (mathematics)outputSequenceInterface (computing)Molekulare BioinformatikCursor (computers)Optical character recognitionPolymorphism (materials science)ACIDLogic synthesisObservational studyRecurrence relationPairwise comparisonOvalLocal GroupControl flowBit rateCommunications protocolCAE Inc.DisintegrationEquals signMetric systemSound effectVector potentialProcess (computing)Group actionFocus (optics)AreaZirkulation <Strömungsmechanik>WebsiteAbstractionTaylor seriesInternetworkingHand fanDirected setMedianCodeSoftware testingSign (mathematics)Integral domainComputer networkDatabaseSequenceCore dumpActive contour modelBitMereologyParticle systemProteinWeb browserCodeResultantBoss CorporationFunctional (mathematics)Multiplication signObservational studySoftwareBuildingGame controllerCommunications protocolGroup actionCellular automatonFamilyDirection (geometry)RepetitionWordNatural languageDifferent (Kate Ryan album)Set (mathematics)Revision controlTask (computing)Mathematical analysisComputer programmingPlanningCASE <Informatik>Data structureDataflowNeuroinformatikDigital electronicsOnline helpSoftware development kitComputer animationMeeting/Interview
Multiplication signLecture/ConferenceComputer animation
Transcript: English(auto-generated)
Thank you for your nice interaction. I'm happy to be here, and I hope I will spread some fascination about life with you. So I conduct, like you heard, bioinformatics researcher.
And bioinformatics basically application on informatics to biology. And it can solve various problems. But in this talk, I will be centered to the analysis of human genomes, so human DNA, in the area of genetic diseases, especially rare diseases.
So for this, some intro to biology and genetics. So what is a human genome? Human genome is the entire set of DNA instructions that are found in a cell.
There is a small mitochondrial DNA. This is not so important, but the most important part is in the chromosomes. There are 23 pairs of chromosomes hidden in the nucleus of the cell. And they hold the whole necessary information
to create the cell and whole organisms. So we have two pairs, which means that one comes from mother and a second one from father. Then the chromosomes, if you unwind them,
there is a double helix structure called a deoxyribonucleic acid. And it's a concept of two strands of nucleobases, let's say letters, A, C, D, G. These letters are paired all together,
always A with T and C with G. So if we know the sequence of one strand, we automatically know the sequence of the second strand. And how many letters are there in our genomes? There is three billion letters. So if we want to store it, we can actually fit it in one CD
if someone still uses it. I don't think so. So instead of that, you can print it out. Not sure if this is a good idea, but actually someone did it before. So this is whole human genome printed in the library.
So one color is one chromosome. And you see the letters. This is really like this. You will not find there anything else than these letters. And these are four, the size is four points.
So it would take really long to read. And, okay, talking about size of a human genome, is it a big genome or not compared to other organisms? We are actually quite advanced and complex. But our genome is not the biggest one, as someone might think. There are other species that have much bigger genomes,
especially the plants and fish or amphibians. How come is this possible? Nothing is that the genome consists of coding part and non-coding parts.
I will explain the coding part, these are genes. And the genes contain the recipes for proteins. The non-coding DNA doesn't hold this information. You might have heard about junk DNA.
It's exactly not the junk DNA, mostly because there are some important functionalities of this DNA, like a regulation of the gene expression, like which genes are switched on and switched off and so on.
So now I told you that there are instructions for creation of the proteins. How do we create a protein from DNA? There is the double-stranded DNA hidden in the nucleus. It's hidden in the nucleus because it needs to be protected,
like so there are not happening, like the changes that we don't want. So it is actually transcribed. There is a copy of the DNA called mRNA, and this travels outside of the nucleus, and there it's translated into protein.
It is translated in a special way because the code, the letters are read by three, like by triplets. So you read UUA, the first letters of the mRNA,
and you will end up with leucine if you look it up in the table. So basically the protein, you can imagine it as a string with beads of different colors. So we created a protein, and the protein is
like the basic structural and functional unit of our cells, and it actually, I told you about the bead, but here is a 3D structure that is somehow like folded, and yeah, that's what the protein does.
And basically the function is hidden like in the structure of the protein. So if the structure of the protein is changed, then the function, its function can be altered. This is very cool.
Maybe you don't see it like it, but these are like real proteins within our cells, and here you can imagine on the left upper corner there is a molecule of water, yeah?
So you can imagine the scale of other proteins basically, and it represents a cell actually. So on the right side here there is a nucleus, and this is the DNA with the proteins that are necessary for transcribing and translating them and maintaining the DNA.
Then we can see a biological membrane here, and there are some proteins inserted into the membranes, and it's like a protein complex. It's not just one protein, but actually you can see these are like protein machines that are tirelessly working
in each of our cells just now. Yeah? If not, then we are dead. So yeah, the good thing is if you go to the website you can really click there and see the real atomic structure of these proteins.
And now I would like to show you how the molecular machines work. So this is a video which shows a white blood cell here
in blue actually, which travels through the bloodstream, and the cell receives a signal that there is some infection in the body. So the cell needs to like prepare and go out to catch the like insider, let's say.
It's not necessary to understand everything. I think it's just fascinating to see it. Like this is a nucleus. These are the building blocks of the cells, the structure,
and the structure is constantly like appearing and disappearing, let's say, assembling and disassembling.
This is a protein complex that inside of this wall, there are other proteins that are carried to another place in the cell.
This is a mRNA released from the nucleus. There's molecular machine that transcribes it into the protein, the green one. Now this is assembled and the yellow thing is protein.
This is other proteins and they are traveling to some other part of the cell.
Now the ball is like merging with other structure of the cell and releasing the proteins to its final direction, escaping the,
it was quite a lot of things that had to happen. So the cell knows and can perform its action. Okay, so this was when everything was working well.
What happens if there is some problem? Actually, as in programming, there are various possibilities. Nothing happens. We have some bug in a comment. It's okay. This can be just variation because everyone writes the comments differently.
So these are just the differences between individuals. Then there can be some problem, which is alternative to disease, or it can be some fatal error, which is a death. So maybe you would imagine that if there is such a big problem
as a death, that you would need some big change in the code, the genetic code, I mean now, in the DNA. But no, one letter change can cause death, actually. How this is possible? If we go back to the genetic code and we are
like performing the, like translating the information, there can happen various scenarios. The effect is no, it's silent. Then it can be missense, where basically the one beat is exchanged.
So there is the proline to trionine. It could be a nonsense variant, let's say mutation in this case, that leads to a premature termination of the protein. So there is part of the protein missing. Or there can be some insertion and deletion,
and it can result actually in alteration of the reading frame where the rest of the protein, it's completely different, something completely else than it's necessary. So we have different variants, and based on its effect,
as you saw that it can be from no effect to very high effect. And we have basically variant frequency. The variants that are like with big effect are usually like by the evolution, like the body is getting rid
of this. And the ones that are not so little, well, the little, we cannot see it in population completely. But if there are some, there are rare diseases. There are also diseases, genetics diseases, that are caused by more genes, like synergy of more genes.
And there are the modifying factors. That's why we are different. So these are not very, these are completely normal. And we have sometimes some beneficial alleles that give some advantage to the organism. So here we will concentrate mostly to the rare diseases,
basically because these are like extreme versions of the mutations, what can happen. And it's very useful to concentrate to them. So talking about rare diseases, what is a rare disease? Rare disease is a disease that affects fewer
than one people in 2000. The one people in 2000, yeah. 72% of them are genetic. 70% of them start in childhood. And there are more than 6,000 different diseases.
But if we take altogether these 6,000 diseases, it affects up to 6% of population. So which means that it's not so rare. And actually here at the conference, I assume there can be like 1,000 people.
So probably or statistically, there could be like between 45 and 59 people affected by some rare disease. And now I'm talking just about rare diseases, not other genetic diseases. So yeah, there are some examples of rare diseases.
Maybe you heard about some. Maybe you noticed this eye bucket challenge. The people did this not because they are stupid, but because they wanted to raise awareness about one rare disease in particular, the ALS, the amelotrophic lateral sclerosis.
Or I don't know in other countries, but here, maybe in the Czech Republic, maybe you could notice some donations organized by the families of spinal muscular atrophy, SMN1, because there was a treatment that wasn't paid,
and they wanted this treatment. So it was in the media. So maybe you heard about some of these diseases. What is it about them? They affect people of all ages, are progressive, they are worsening over time and leading to death,
potentially, not all of them. They can affect any part of the body and have various symptoms, physical, neurological, developmental or behavioral. And what is true for many of them, the genetic cause is not known.
The patients are very often going from one doctor to another, and they are undiagnosed for years or even decades or for whole life. So this has a big social and emotional impact to them and their families. So why should we study the rare diseases?
One thing is to help the patients to treat them, because the thing is that there is treatment just for 5% of the diseases. And the diagnosis is important usually for successful treatment, because many of them
you cannot reverse what was already broken in the body. So then it's for family planning and also the people can have a peace of mind if they know what's happening. The thing is I already said that diseases are not so rare and important thing, the rare diseases are great models
for studying cell physiology and photophysiology, and they can provide us with new insights into biological function of disease-causing gene, which help us to better understand the complex diseases. The thing is that there are between 20, 25,000 of genes,
but the function is not known for many of them, so there are still a lot of work to do. So there are various rare diseases, but a good thing is the approach how to study them is the same for all of them,
or can be there are some variations, but basically what we need to do is to read the DNA, read the important letters, let's say, and find the change of the letter or some other alteration. So how do we read the DNA?
And how do we read human genome? Actually, anyone knows when was the first human genome sequenced? Some guesses? No? Okay, so it was in 2001, actually.
So this is when we, for first time, could read the whole genetic information. And the project took 14 years, and it cost $3 billion, so it was really expensive. The price of the sequencing of human genome was dropping,
and it was following the Moore's Law. And here was the change in 2008, and the change is because there were next generation sequencing technology available,
and it was spread, and all together with knowing the reference genome, it helped us to sequence for a reduced price, and now basically we are at $500 per human genome. So it's more accessible.
So I mentioned two different technologies. There is the Sanger sequencing, the Fisherman, that was used for sequencing the first human genome, and he is able to catch just, let's say, one gene. And then we have the NGS sequencing, and this method is capable of taking
all of the genes from the pool and sequence them at once. So this is a very big advantage that we have. How does it work? We collect biological samples, bugs, blood, or swabs, the buccal swabs.
Then we extract the DNA, and we cut it into small pieces. We take these small pieces and put them to the sequencer, and we receive the FASTQ file, which contains the sequences. And the sequences are around 100 base pairs
with the second generation of sequencers. So imagine that we take the library and cut it in 100 letters, what a mess. So this is the Illumina sequencer. This is where you put the DNA on this last slide,
and then the machine gives you the sequence. Imagine if we would be like James Bond, who was able to sequence DNA on way
because he had this cool device that could insert into the computer, pour the DNA into it, and just read it in real time. So well, we are there. It's me in 2017 using this device.
And yeah, it doesn't read so many sequences, but there is other advantages. It's called Pinion, and the technology is called nanopore sequencing. What is the difference then from the previous one? The difference is that there is no fragmentation.
It's capable of reading the long molecules of a length like kilobases or megabases, like number of letters. Base pair means how many letters you can read, basically. So it is very cool.
This is how it works. This is a biological membrane, basically similar one like you saw in the video here. And there is a protein machine, nanomachine, that takes the DNA and translocates it through a spore,
and there is like electric current, and basically we are measuring the changes in the electric current as the DNA is passing, and every letter produces a different change. So this is how we can read it.
So this is great technology, which is also the advantage that you can just take back and go anywhere to sequence. Anyone, was this sequencing performed? Do you know? Yeah, exactly. Because the thing is that you cannot take the DNA,
if you want to study what happening with the DNA, you cannot take it back because it got changed, basically, during the space traveling. So now we have the sequences. We prepared a sample, we sequenced it,
and we have the sequences. So what do we do now? So we can identify the disease-causing genes and make new discoveries. The files, the raw files, basically, are quite big.
For a whole genome, it's 150 gigabyte basis. Why, if the genome size is this? Because you need to read the genome more times because there are errors in the readings. So basically you need to compensate and with the, like, reading it more,
more times, so it's quite big. And there is another application called whole exome sequencing, and this is basically 2% of the genome, and these are the genes. Actually, genes are just 2% of our genomes.
So this is very convenient way to sequence because it's cheaper, the analysis is easier, and it's very probable that the disease-causing mutation is lying within this region. And the targeted sequencing, you can also choose just some genes, not all of the genes. Let's say you are studying heart diseases,
so you just say you can choose just some genes that are known to be causing the heart diseases. Yeah, so actually I had a, there was a funny because a friend of mine told me,
oh, I have my genome sequenced. Can I send it to you? I said yes, and she sent it to me by email. So, okay, so it wasn't a genome. It was just couple of variants, actually, and maybe you come across this as like the, it's even less genes than the targeted sequencing.
It's like the Ancestry and MyHeritage, for example. Maybe you did it, some of you. So yeah, this is not a whole genome. Maybe you would like to see how the raw sequences are,
what is the data format, let's say. So it's, the sequences can be stored in the FASTA format, the first one. This is how the reference sequence is stored, for example. For the sequences that we receive from a sequencer, it's a FASTQ, the Q stands for quality, and basically it's because there is information
about how quality, which quality has this letter, how we can rely on that it's correctly read. So this is what we are using, and now we would basically, yeah, as you imagine the library in 100 letters,
cut it into 100 letter pieces. So we want to know from where every fragment or every, the sequence came. So this is a read mapping. It's like a puzzle, you can imagine a puzzle,
where there are some missing pieces. There are some broken pieces. These are the sequencing errors. There are duplicated pieces. These are repetitions. Our genome is full of repetitions. And there can be also some contamination. So basically, we, this is the reference genome.
You can imagine it as the box of the, like the picture that you have for the puzzle, and you are trying to find the place of each of the sequence. This is the result. This is BEM file.
This is binary form of the SEM file, the sequence alignment map. And basically, what is important here, it's the sequence that we were matching to the reference. Here is an information about where it goes to the reference. It starts at position seven and goes to the position 30.
And this is like encoding if the letter is matching the reference or there is some insertion or some deletion. So this is how it's encoded. If you visualize it in a genome browser, it looks like this.
And here you can nicely see also the difference between genome sequencing and exome sequencing where there are just the genes and the rest is not covered. So now we are interested in the variants. So variants, we're basically looking for the places
where the sequence of the person is missing, like different from the reference sequence. So here we have the variation at this position and it has different letters for mother and father.
This is the variant call format. This is just for your reference. This is like a list of the variants. Now what we do with the variants, it's like finding a needle in a stack of needle because it's just list. So basically what we do is we have the patient sequence
compared to reference and we want to know what the rest of the population has at these positions. And we are interested in rare variants that are not present in the population. Then we want to know something about the sequences,
about the variants, sorry. So we remove the low quality ones and then we annotate them with information from various biological databases. There are a lot of them and contain different information about frequency, about their biological functions, about where they are in the cell.
So many things like that. And then you are trying to filter based on this information and find the one variant. Here is it. How long actually does it take to perform the sequencing experiment for genome, exome or targeted sequencing?
So depending on a machine that you have because you cannot analyze genome on your laptop, this is in the CPU hours. So you can calculate how much would it take on your HPC or what you need
as a specification for the cloud machine. So here is a summary of the bioinformatic workflow. If we are lucky, we are ahead of the candidate variant. If we have the variant, it's a good idea to make a validation. For this, we call the fishermen
because we already know which gene it is located in and we make sure that it's real. Then we check the family if all people with the variant are ill. And then it's good to find another family, not non-relative family
and check if the variant is also there. And if we don't know more, then we need to perform studies in vitro and in vivo and to learn about if this variant is really producing this,
let's say the picture of the disease in the cells. So if the variant is confirmed as a disease causing, yeah, so what do we do? If the treatment is there, that's cool. It's not very likely, as I told you before, unfortunately.
But maybe there are some drugs that affect the same metabolic pathway. And if we now understand how it works, basically the gene, then we can find some substance and repurpose this medicament for another disease.
If there's no treatment, we at least can help the family with genetic counseling and family planning so we can avoid that they have another kit with the disease, for example. And then it's good to let the world know about it
if it's something not very widely known. So publish, submit it to databases. And if you need to perform some studies, then you can connect with other researchers. There is a gene matcher, for example, which is, someone told me that it's like a tinder for biologists.
So basically the interest in the same gene, yeah. So then you, yeah. If the variant is not identified, what do we do? We will check if it's well sequenced because sometimes the quality of sequencing is not very good.
So if not, we sequence again. Then we can also reanalyze it with different tools, settings, also there are different versions of reference. And so we can also wait and reanalyze it later with the novel findings. It many times helps.
We can also try the long reads, the cool nanopore sequencing. And also there's a possibility that it's not genetic. The good thing is that thanks to the NGS sequencing, the clinical diagnosis improved from 1%,
that was when just the fishermen was there, to 50%, which is a great hope. Yeah, but what about the python, right? So in this bioinformatic workflow, the python, many tools, basically in the bioinformatics, usually there are some tools for the,
like let's say standard tasks, that you just put them all together into some pipelines. Some of them are written in python already. And also there is the Bioconda, which is a channel for Anaconda, where you can install a lot of biological packages
that are important for the NGS analysis. So yeah, then as you are piping the programs all together, there is a snake make, which is written directly in python that can help you with this task,
or next flow or riddle. The cool thing about next flow is there is an NF core, a portal, where are already done, pipelines that you can just use. So this is a great thing.
So basically or along there is the python. There is also a biopython, which is a tool for computational molecular biology, and it lets you manipulate the sequences. It allows you to like somehow compare them all together, and also do some population genetics,
structural bioinformatics, and also connect with the biological databases. Yeah, here is just an example, funny quote about as I told you how to get from DNA to protein. So this is done in biopython. And then important for the analytical part,
where you are making sense of the variants, python is very useful because in many labs they are using just Excel to filter the sequences and stuff like that. So yeah, for this pandas, Jupyter, NumPy is very cool.
Or there is also a package circuit bio. I don't have personal experience with this one, but it also looked cool. Okay, and one case study at the end. So the patients, it was a healthy Caucasian female, and she experienced three miscarriages
at eight or 10 weeks of pregnancy. She underwent many examinations, and the only thing that was found was like in her sputum blanks, there was detected the meningococcal streptococcal infection, nothing more.
She was treated with antibiotics, but other things, she was healthy. So she decided to get her genome sequence, and there was one gene, MTHFR mutation,
homozygous mutation. And basically this mutation reduces the enzyme activity to 30%. This is, you see the real mutation in the genome browser.
This is the pathway that it affects. Basically, this gene acts here and converts the folic acid to another substance, to 5-MTHF. So yeah, the enzyme function is reduced
and in the literature, PubMed, there's a PubMed, so she performed the PubMed search actually and found a treatment where they were comparing two different treatments for this particular mutation. And the treatment worked very well. It was almost the same as in the control group.
So she underwent the treatment with a bit-modified protocol basically instead of the folic acid, she took the later one, the 5-MTHF, because there was another study about it. And she was successful.
Yeah. So yeah, this wasn't a rare disease, but it's like actually there's another rare disease
of this same gene that by understanding function of this gene, it has implications for many of us. And the thing is that it takes from 15 to 17 years to get the knowledge found in our research
to a clinical practice. It's a long time to go and yeah, it takes time. And actually there is like in the both of my lab did like retrospectively looked at how the patients get,
the kidney patients get there. And 25% of the families diagnosed with the kidney disease were like a result of direct family referral and basically the pediatricians didn't like help them to find the center and to know the disease.
So actively pursuing cell diagnosis can actually be useful. There are a lot of networks that can help you with this and you can find the place probably where they are specialized to the condition. Okay, so thank you.
Okay, thank you Anna, it was fascinating. Thank you very much. Because of the technical issues, we took all the time so we don't have time for questions, but feel free to find Anna on the hallway
and stick here because in five minutes we'll host lightning talks in this hall. Thank you Anna, thank you again. Thank you.