 So welcome back, everyone, for the second part of this lecture. I hope this is clear. So before the break, I quickly try to explain to you guys what hemizygous recessive is and why it is important when we talk about disease. Because a lot of, well, not a lot of disease, but some diseases are located on the X chromosome, meaning that as a male, if you get the X chromosome, which is broken from your mother, and then you will always pass it along. While if you are a female, because you have two X chromosomes, you might be lucky, and only one of them is broken. So then you have a 50% chance of passing it on to your offspring, which really makes a difference in the inheritance pattern. And when we start looking at inheritance diagrams later on, we can see that it makes a real difference sometimes if it comes from males or females. Although that won't be this lecture. That will be a later lecture. OK, so is there any questions about the first part of the lecture? I did get a text saying that did we already do the assignments from last time? And I will make some room at the end of the lecture for people who have questions about the assignments from lecture one, because I looked at my GitHub and I saw that only four people were able to make a fork of my repository. And I think two out of four were actually successful in updating their version. So it would be good that everyone at least tries to get the version control and the GitHub thing working, because it is an important part in bioinformatics to be able to do reproducible research. Because software changes, so you need to be able to go back into time to redo an analysis as if it were 2019. All right, so let's just continue. So next slide. Here we see an example of an inheritance diagram. So here, this is from Morgan. It's a kind of restyled one, because the original one was drawn by hand. But here we see the experiment that we just showed, right? So the experiment that we just showed was we have females who have white eyes and miniature wings. So they on one chromosome have a W and an M gene. And on the other chromosome, they also have a small W and a small M. And of course, we have wild type males. So the males have a W plus a Leo, because they are wild types, so they have a normal eyes and they have normal wings. And of course, we have this little bar underneath. And this means that there's the Y chromosome, so they are hemizagous, so they only have one copy. So that's the way that you would write it down in genetics. So then when they did this cross, then they observed indeed that the individuals coming out were wild type females. But all of the males had white eyes and miniature wings. And that is, of course, because every male gets their X chromosome from the female. So all of the males get a W and an M, and they have a normal one. Well, here the females, they get a WM chromosome from the mother, but they always get the W plus M plus chromosome from the father. And since this is a dominant phenotype, they will look like wild type Drosophila. So what happens now when we start crossing these animals? So this is the F1 generation. So if we take these wild type females, who are not really wild type, right, because they have two different copies, it's just that both phenotypes are dominant. So when we cross these wild type females to the males that we get, we get the F2 generation. So in the F2 generation, something interesting happens because now we see all of a sudden an explosion of variants, right? So we see that we have white-eyed miniature wings, individuals which are female. We see that we have males again, which have white eyes and miniature wings. We see that we have wild type females, wild type males. We see that we have white-eyed females who have normal wings and we see that we have males who look the same. So they have white eyes, normal wings and we see that there are the other situation where we have wild type eyes and miniature wings, right? So there is some breaking up going on and this is due to recombination, right? So now we can just start counting, right? Because in this case, we can define if the animals look like the parents, right? Because these look like the father and these look like the mother. So we call them parental phenotypes. We also have animals which have one feature of one parent and a feature of the other parent, right? So they have recombined. So we can define these as the recombinant phenotypes, right? So just looking at the animals, we just classify them based on their phenotype. So when we add up the numbers here, right? And these numbers are just an example. I think they are actually the real numbers from the original experiment. But what we observe is that in total, we have 750 animals who have white eyes, miniature wings. We see 791 who are wild type. Then we have individuals who have recombined and we see other individuals which have recombined the other way, right? So in total, we observe that we have 1541 offspring of this cross who has parental phenotypes and there are 900 which have recombinant phenotypes. So now we can actually start calculating the distance between these two phenotypes, right? On the X chromosome. So how do we do that? Well, first off, we calculate how many offspring we had in total and then we just say, well, 900 out of the total recombined and then we multiply by 100 and then we see that we have 36.9 centimorgan between the W gene and the M gene. So both of these genes located on X, the distance between them is 36.9 centimorgan, right? Because we multiply by 100, so we have centimorgan. It's actually 0.369 Morgan. But that's why we use centimorgan because then we get into a percentage, right? So this is the way that we use to build genetic maps before we had any idea of the genome. So in the F2, the most frequent phenotypes for both sexes were the phenotypes of the parents in the original cross. So wide-eyed, miniature wings, red-eyes, normal wings, right? So that's this group here, right? So these four types that can occur. Nonparental phenotypes, wide-eyes with normal wings or red-eyes with miniature wings occurred in about 37% of the F2 flies. And this is well below the 50%, which is predicted for independent assortment because if these things would be independent, right? If the two genes would be located on different chromosomes, then we would have expected half of the individuals to be recombinant, right? Because the genes would not have been located on the same chromosome. But since they are linked, right? We see that 37% has the nonparental phenotypes, while 63% has the parental phenotypes, we can now say that indeed these two genes are located on the same chromosome and they have a distance of 36, 37 centimorgans between them, right? So how did, because Morgan just had this observation and now he needed to come up with something like how does this happen then, right? So what he proposed is that during meiosis or during when sperm and axels are created, alleles of some genes assort together because they are near each other on the same chromosome. So that's where his theory of chromosomes come in. He then proposed that there is something like recombination, right? So had parts of the X chromosome on the mother genome are exchanged with the other X chromosome in the mother. And this does not happen in the father, right? Because the father only has one X chromosome, it cannot exchange genetic material or it cannot recombine the two chromosomes together, right? So crossover occurs at the four chromatid stage of prophase one and meiosis. We know this now, of course, like Thomas Morgan Hunt had no idea about this and all chromatids may be involved in crossing over as chiasmas form along the aligned chromosome. So that is just how it works in real life, right? So if we get to the four stage, and I think I have a picture of that. So let's forget about this last statement for now. But this is the invention that Thomas and Morgan made. So how does this work? So just as a different representation. So if we have two traits, which are Mendelian, they can be very far apart on the chromosome, right? So there is a low chance that both phenotypes will be passed on to the offspring. If we just talk about the X chromosome, right? So we call this, they are weakly linked together. You can also have the situation, of course, where you have two Mendelian traits which are very close together on the chromosome. And then of course there's a high chance that both offspring will get, or that the offspring will get both phenotypes, right? So that is the concept of linkage and everything in genetics is based on how strongly linked two things are. And of course we're talking here about phenotypes and that's because it's the phenotype lecture, but this also holds for genetic markers. If you can measure parts of the genome, either using SNPs or AFLP markers, then also two markers on the same chromosome will be linked together. And if these two markers are at the end of the chromosomes or very far apart, then their linkage will be low. If they are more or less very close to each other, then their linkage, so the correlation will be high. Okay, so that's the one point cross. This was normally the break for the slide at which I would break for break. All right, so we call this a two point cross, right? Because we have two phenotypes, so we are looking at two points in the genome or two genes in the genome and this type of a two point cross can be used to determine if genes are linked. So if they are connected to each other on the same chromosome or if they are independent, which means that one of the genes is on one chromosome and the other genes is on the other one. But like we saw, we can also see that it can be used to estimate the distance between two genes. So had the typical setup for what we showed, right? Here we showed the Morgan original experiment. And here this is based on the fact that males actually only have one copy and females have two copies. But we can do a two point cross as well for phenotypes which are on the autosomes. So it's not only sex-linked phenotypes, we can do the same thing for the autosomes. So what do we do then? Well, in a two point cross, we take an individual which is heterozygous for phenotype A, we take an individual which is heterozygous for phenotype B and then cross it with an individual which is homozygous for A and homozygous for B. So it has two small A's, two small B's, while the other individual, so the mother has, is a heterozygous, while the father is homozygous for both phenotypes. And then we set up the two point cross, so we make like before, so we made them together, we get the F1 hybrids, right? So the F1 hybrids, they have a mixture of the two phenotypes and then we cross these two hybrids with each other, generating the F2 population and then we use the observed parental phenotypes divide, or we use the observed amount of parental phenotypes versus the observed amount of recombinant phenotypes. So what do we calculate? We calculate then the amount of recombinants divided by the total amount and that is the genetic distance. So 17 map units in this case. So you can look at the cross here in more detail. But it follows the exact same strategy as the original cross for Morgan, but of course, Morgan didn't use a homozygous individual, Morgan exploited the fact that his two phenotypes were located on the X chromosome. So two point cross allows you to do two things, determine if a gene is on, or if two genes are on the same chromosome or if they're on different chromosomes. So, and when we talk about genes in point crosses, we actually mean phenotypes, right? Because we are only talking here about Mendelian phenotypes. So the phenotypic effect, there has to be an observable difference between individuals having AA versus individuals who are small A, small A and this difference need to be clear, right? If there's no real difference in the phenotype, so to speak, then we cannot do this. So here actually we see a three point cross, right? So it's the same structure again. So we have one individual, so here we're not just having two, but we're having three. So it's the same thing again, right? So we're just crossing them, then you cross the offspring together, you generate the F2 population, you see how many there are, which have phenotypic differences observable as the father or the mother. But in this case, of course, we cannot see the difference. So it's just a little example, but so we are talking about phenotypes, these phenotypes need to be Mendelian and the difference needs to be observable. So either qualitative or quantitative, we need to be able to measure this distance. And that is very important when you start doing this mapping using phenotypes. So I just showed you guys this. So just to show it a little bit bigger. So here again, we have three. So we have three different genes that we're looking at or two in this, no, two in this case. But in the end, you just determine how many parental phenotypes you have, you determine how many recombinant phenotypes you have, and then you just calculate the number of recombinants divided by the total amount of offspring that you generated. And then based on this, you can say how far apart two genes are. So we'll just look at the PowerPoint later on and try to figure it out. You can also make the crossing scheme, of course. So besides the two-point cross, what we also use a lot is a three-point cross. So in a three-point cross, we have three genes instead of two. So this can also be used to determine if these three genes of interest are linked or if they are independent, right? If all three of them are in different chromosomes or if two of them are in one chromosome and the other is on another chromosome. But if all three of these Mendelian phenotypes that we use for mapping, right? If they are all three on the same chromosome, then we get the distances, right? So we can calculate the distance from A to B, we can calculate the distance from B to C, and we can calculate the distance from A to C. And by knowing these three distance measurements, we can also infer the order of these genes on the chromosome. So if it's A, B in the middle, C at the end, if it's B at the beginning, A in the middle, C at the end, or if it is A at the beginning, C in the middle, and B at the end, right? So we can also make an ordering. So we can start building up a map, a genetic map. So using a three-point cross, you can determine not just the distances, but because you know three distances, you can also infer the order, right? So in a three-point cross, so typically geneticists design experiments to gather data on several traits in one test cross and an example of a three-point cross would be where we have an individual, which is wild type P, wild type R, wild type J, recombinant P, recombinant R, not wild type, but like another allele of P, another allele of R, and another allele of J, and we cross this with an individual, which is homozygous recombinant, more or less, which has the different alleles, right? So it's just the wild type allele, versus the wild type allele, versus the mutant alleles, and then this individual is homozygous for the mutant alleles, right? So in the progeny, each gene has two possible phenotypes. Have for three genes, there are two to the power of three is eight different expected phenotypic classes in the progeny, if they are not too heavily linked, right? If two genes are 100% linked, then of course there's only four, but has so the two to the power of three. So how does this look? So this is a three point cross, right? So here we have the genes that we're talking about. So here, and I like this representation a lot because I think it's very clear. So we can see here based on the, so we have two chromosomes from the first parent and we have the second parent, and we can see here that we have the wild type allele, we have the mutant allele using different colors, and here we see the same thing again, but this individual is homozygous for the two mutant alleles. So here the phenotypes that we're looking at are phenotypes which were studied by Gregor Mendel. So you are yellow, you will have an elongated fruit, or you, and you are dry, right? So you're not juicy. And here, parent two is purple, round and juicy, right? So here, of course, the yellow phenotype is dominant over the purple phenotype. The elongated phenotypes are being long is dominant over the round phenotype and the dry phenotype is also dominant over the juicy phenotype, right? And we know this because if that would not be the case, right? If it would be an additive effect, then this individual would not be yellow but it would be a mix between yellow and purple. It would not be elongated. It would be a mix between the two. So three point crosses are always done on dominant phenotypes and never on additive phenotypes because we cannot decompose the mixture, right? So, but when we do the cross, so we take the first cross, so we test parent one, we cross it with parent two, we get the F1 generation, we then cross the individuals of the F1 generation with each other, and then in the end, we have eight different types of phenotypes that can come out, right? So this is just an example. And so here we see the wild type, right? So it's yellow elongated and dry. We have the mutant, which is purple, juicy and round, and then we can have the different combinations, right? And we just have, so in this case, we made 500 offspring from the F1 and now we can just count the number of offspring of each of the different types that we observe. And then based on that, we can then figure out where these genes are located because here we can figure out the number of recombinants for the A and B phenotype, the number of recombinants for A and C, the number of recombinants for B and C, and then just divide that by the total number of offspring every time, right? So the recombination frequency is the ratio of non-parental phenotypes to the total individuals. It is expressed as a percentage, which is equivalent to the number of map units between two genes, and so as an example, if a hundred out of a thousand individuals display the phenotypes resulting from a crossover between genes A and B, then the recombination frequency is 10%, or we can also say that genes A and B are 10 map units apart on the chromosome. And then they did this, right? So we had drosophila and they started collecting all kinds of mutant drosophila. So we have ones which have like standard antennas versus small antennas, normal wings versus miniature wings, big legs versus small legs, different colors of the skin, different types of eyes, so not just, so having red eyes and purple eyes. So all of these mutants were wild mutants from drosophila that they caught and these were more or less, these were all studies. So they all did crosses between these different ones to measure the distance. And then what happens is that they determined the distance between these two phenotypic markers to be like 13 centimorgans. And then when you do a three point cross you can determine, okay, so the yellow phenotype is 31 units away from the red phenotype, the orange one is 13 units away and this one is then 18 units away from that one. And if you know these different distances then you can start building up a genetic map. And this is the genetic map of drosophila that was created in 1919. So before we knew anything about DNA, before we knew anything about how inheritance is really working in real life, Thomas, I have Thomas Morgan Hunt or Thomas Hunt Morgan was already able to determine that drosophila has four chromosomes. He was able to determine the length of each of the chromosomes and he was able to build up a genetic map to allow to associate regions of the genome with certain phenotypes. And these genetic maps were built up of purely phenotypic observations of dominant phenotype. And it is being said that the Morgan's theory of the chromosome is a great leap of imagination which is comparable to what Galileo or Newton did because it took us out of the dark ages where we knew nothing about inheritance or how inheritance worked into a world where all of a sudden we can determine how many chromosomes there are, how long the chromosomes are, which gene is located on which chromosome and this map that they made in 1917 is still 100% accurate today. So even though we now sequencing technology, we have all kinds of like massively expensive equipment, the work that they did by just crossing drosophila flies together is still valid today. And we now know exactly which genes are responsible for these phenotypes that we're looking at. So we know the exact gene on the genome, but the work, so to go from having this idea like oh, so we observe that phenotypes are inherited together, which is not exactly how Gregor Mendel told us. This is like in genetics, this is like the major step forward. This is where all of genetics is built on, all of the last 100 years of progress which we made in our medical field is based on this theory because without this theory, we would still be kind of struggling in the dark, not understanding why some people have diseases, have things like Huntington and all of these diseases, they are studyable by the theory of chromosomes. So today, of course, we have different types of maps. We have high resolution maps which includes both genetic markers from test crosses. So hey, if you look at the genetic map from drosophila, still the sapia eyes and the hairy body, they are still genetic markers. We still use them nowadays, but we have complimented this map by all kinds of DNA markers. So DNA technology where we use PCR or we use sequencing to determine SNPs or other things in the genome which we can use to distinguish individual, right? So, but genomic sequencing allows to determine of exact positions of genes and these physical maps use, of course, molecular tools rather than data from crossover studies. So, but so nowadays we have SNP chips where we genotype 50,000 positions in the genome, but the whole theory of genetics is way, way, way before we knew anything about how DNA worked. And I still think that that's amazing. Like I love telling this story and I love kind of getting people excited about the fact that you can know stuff and know how stuff works without even knowing what the carrier is. And without knowing that DNA existed, you could already predict in 1920 how certain fruit flies would look based on just crossing them together and you could build up a whole genetic map to allow to associate regions of the genome which they didn't know was encoded into DNA but associate regions with the genome with other phenotypes. All right, so that's all I wanted to say about Mendelian phenotypes. So, Mendelian phenotypes are markers. You can build up a chromosomal map using dominant Mendelian phenotypes and it's just amazing what you can do with them. So, if we talk about complex phenotypes, we have differences in many genes that cause the difference in the phenotype that is observed between individuals. Some example of complex phenotypes, almost all phenotypes that we know of are complex except for the ones that are Mendelian but things which we know are very, very complex in a way is human stature. Like there's literally hundreds of genes which are involved in how big you are. Obesity, right, how fat you are or how you respond to certain foods head that is not controlled by a single gene but by many, many different genes. The flowering times in plants is actually a complex phenotype. A lot of people used to think that flowering time was more or less Mendelian because of experiments done with Arabidopsis and the false understanding came from the FTO locus. So there's one locus in the, I think chromosome five of Arabidopsis and that is a major modifier of flowering time. So if you have the working allele, you flower probably like 20 days earlier than when you have the broken allele. The flowering time itself is not just controlled by this one FTO locus. There's many different modifiers. So also there, there's probably like 50 to 100 genes involved to control when a flower or when a plant starts to flower and after being planted. Another very common example that's being used is milk yield, so milk yield in cows. Again, there's, as far as we know now, there's one very major modifier which kind of gives you plus 20% milk if you have the right allele and if you have the wrong allele it's like minus 20% milk but of course there's literally probably hundreds of genes in the genome which control how much milk a cow gives. So when we talk about complex phenotypes, we are always interested in finding out the genetic architecture of this phenotype. Which genes are involved? How much do they contribute to the phenotype? Do they make you bigger? Do they make you smaller? How much do they make you bigger and smaller? And nowadays we use QTL mapping to study inbred populations to study these complex traits. So we have a genetic map, we have our phenotype and then we try to find out where in the genome are the locations or where in the genome are genes that are controlling my phenotype. And if you do that in an inbred population, right? So in a population where you control who breeds with whom, then we call this QTL mapping. And if we use a natural population like a human population where humans are free to choose who to mate with, then we call this genome wide association. And we will get back to this, we will do QTL mapping and I will try to teach you guys how to do genome wide association. So to find which part of the genome is involved in regulating a certain phenotypic observation. And that is actually kind of the biggest part of my work. So the biggest part of my work involves doing experiments and we mostly do experiments on, so we do QTL mapping when we are using our own mouse population. So we do crosses between mice, we generate populations and then we measure phenotypes and then we try to determine which part of the genome is influencing our phenotypes. And we also do genome wide association in our group here because we are also very interested in milk yield and cows, so we have done and we have published a lot of studies where we literally measure like hundreds to thousands of cows and try to figure out where the milk yield is controlled, which genes are involved in the fat percentage or the protein percentage. So that is a big part of my job as a bioinformatician. All right, so we're a little bit out of sync with where I had the breaks. So before the break, I told you about Mendelian and complex phenotypes and about linkage and genetic maps that you can make from Mendelian phenotypes. And after the break, which we won't have a break for another 15 minutes, so we'll just continue, but I will talk to you about databases with phenotypic information and about how we do statistical analysis of phenotypes. Good, any questions so far, by the way, because this is a slide where I normally just wait for you guys and like, do you have any questions about a two-point cross, a three-point cross about phenotypes in general, about Mendelian phenotypes, about complex phenotypes? Then we can take like a little bit of time to just discuss it, talk about it. So if you have a question or an observation or just I didn't understand anything from two-point crosses and I do wanna know how a two-point cross works more in detail, then we can do that. Although the assignments for today will also involve you doing a, well, not doing a two-point cross, right, but like looking into the data of a two-point cross to kind of figure out if you can calculate the distance between two genes based on observations. Good, no questions. That either means that everyone's asleep or it means that everyone's already like tuned out or everyone understood and I did a good job explaining it, but I never assume it's the third one. I always assume that either everyone tuned out already or no one's listening anymore and I'm just talking to myself, which is fine. I do that sometimes as well. All right, no questions, we just continue. All right, so about databases, you're doing a good job, yeah, well, thanks moderator. Hey, you got a different icon, what happened? Oh my God, you subscribe with your prime account. That's so nice that you got prime. Anyway, so what is a database? So a database is an organized collection of data, which is logical. It is the collection of schemas, tables, queries, reports, views and other objects. So physically, of course, a database server is a dedicated computer that holds the actual database and it's just running database software and all of this related software, right? So a database itself is defined as schemas, so schemas determine what is stored in which column and what kind of type you store in there. So this is a column which holds dates and this is a column where you have user names and this is a column where there's passwords in there. Then you have tables, so tables are like Excel tables with the columns according to the schema. A query is something which you can ask a database and then you have reports, that's the answer that the database gives you and then a view is more or less a subset of one of the tables. So that is what you are presenting to the user. So the phenotypic databases that I actually wanted to talk to you about is a couple of them which I think are very important and like I told you at the beginning, it is a little bit focused on mice because I do experiments with mouse and of course, we also do experiments with cows and other animals. But for me, I just wanna talk to you guys about my work. So IMPC is the International Mouse Phenotypical Sorciam. So the link is here. So if you get the PDF from Moodle, you can just click on them. I wanted to show you the OMIM database which is the online Mendelian inheritance in men. So that is for humans, which Mendelian phenotypes are there? Where are they located and what are the effects of them? The Gen2Fend database, which is the genotype to phenotype database, it is the database which contains a lot of QTL information. And I also wanted to talk to you guys a little bit about Genetwork which is a database which contains a lot of data over the last 25 years. It's phenotype data on mouse, but they also have phenotype data on other species. So it's not just mouse, they also have barley and a couple of other species in there. But it's originally a mouse database for the mouse community. So there's a lot of data based on different mouse inbred strains and crosses between these inbred strains. All right, so IMPC, this is how the website looks like. We will go to the website as well. So actually let me open up a Firefox window for you guys so that we can go to IMPC. But I have another slide. So IMPC is a very interesting database because they produce germline transmissions of targeted knockout mutation in embryonic strand cells for 20,000 known and predicted gene in mice. So essentially what they do is they started at the first gene on chromosome one, made a knockout mouse, and then made children of those knockout mice and then see if anything happens to the phenotype of the mouse. Then they moved on to the second gene on chromosome one, knocked it out, saw if there's any phenotypic difference, right? So their goal is to knock out every gene that is known or predicted to exist in mouse to kind of get an idea what every gene does in the mouse genome, right? And of course, if a mouse gene is knocked out and the mouse gets a certain disease, then of course the homologous gene in human is probably also causing that disease in human. So it's a big fishing expedition and have one by one they are knocking out genes and they are trying to get kind of the idea or they try to get an annotation for every gene in the genome. So how did they do this? Well, they test each of these mutant mouse lines through a broad-based phenotyping pipeline. And so they look at all major adult organ systems and most areas of major human diseases. So these mice are more or less auto-phenotyped through a massive system of phenotyping. And the nice thing is they provide a centralized data center so they have a big database and all of this data is available for free, which is just amazing to think about. Had that amount of money to just make a single knockout mice, well, it's not that expensive anymore with CRISPR-Cas, but like 10 years ago when we didn't have CRISPR-Cas and we had to do germline transmissions, right? So we used E and U mutants. It was really expensive. So getting a knockout mouse would be like 10 to $15,000. Now a days you can get a knockout mouse for like $2,000. But if you multiply that by 20,000 genes, then we're talking about a database with information which is worth at least 20 million to 40 million euros. And that is insane that they are just offering this for free. So let's take a look at the website. Let me push a button and then you should be able to see the website. So as you can see here, they have 7,824 genes knocked out today, right? If you would have looked like three years ago, it would say like 4,000. And it's going faster and faster because of new technologies, because of CRISPR-Cas, it's easier and easier to create knockout mice. So, and each of these mice have been phenotype. All right, so question to you guys. What's your favorite gene or your favorite phenotype? Because they allow two ways of searching. So you can search by genes or you can search by phenotype. So is there anything that you guys are interested in? Is there any gene that you are currently studying or that you think, oh, that's an interesting gene? Just give me an example. I can fill in my favorite gene, but that's a very boring gene. I don't even think that they have a knockout mouse for that because when you knock out the gene, it turns out that mice don't get born. So, that's my favorite gene. But I wanna get just one of you guys, so we can throw it in. And if you have a favorite phenotype in mouse, right? I really wanna know what is controlling tail length in the mouse or I want to see eye color in the mouse, then you can just search for the phenotypes. So just so that you guys get an idea and we can just talk you guys through the different screens that they have, feeling that everyone's sleeping again. It's a little bit of a shame. Come on, people, give me a gene. What is your favorite gene? As a biologist, you should have a favorite gene. Everyone should have a favorite gene. Go for obese, all right, we go for obese. So obese is of course not a gene, but a phenotype. So we can just say, well, obesity. Right, so when you search for obesity, it says, well, we don't know exactly what obesity is because obesity is in this case defined as abnormal body weight, right? So the body weight could be lower, it could be higher. And of course, it knows that obesity is also related to things like body size. I wasn't sleeping, but never thinking about my favorite gene. Well, you can just pick one, right? There's 20,000, so the chances of two people picking the same favorite gene are relatively small. So there's, like you can see that they use ontologies. Like we talked about gene ontology, there is also here phenotype ontology. So obesity is itself is not a official structured ontology, right, a vocabulary which is fixed so that you can always talk about the same thing. So, but hey, if we look at abnormal body weight, it says that there's 144 genes associated with this phenotype, meaning that from the 7,000 something knockouts that they made, 144 of these showed a difference in the body weight. So in total, they actually tested 8,185 apparently. So hey, and then you can see here, okay. So for example, the act two gene, when it is homozygous in males and in females, when they are early adults, when you knock it out, you decrease the body weight. So, and this has a very, very significant p-value of zero, which means that it's like definitely for sure that this gene is controlling the body weight, right? So, and if we want to know more about this gene, we can just click on it, right? Then you go to the act two data chart, right? So here we go, of course, because we searched in relationship to body weight and body weight, we already go to the body weight table. So what do we see here? We see that body weight was performed on 445 mice. The chart showed the result of body weights in 25 females and 17 male mutants, compared to 177 female and 200 controls, right? So just to give you an idea of the amount of data that is in there, literally thousands and thousands of mice are or less studied for this project to give us an idea, right? So what do we see here? We see that there's a testing protocol that they used, a certain testing environment, right? So that's all standardized between different laboratories. The background strain, so there are different types of inbred mice, so they use a certain strain and then compare the knockout to the standard strain. And of course it says decreased body weight, which we already knew. So what you see here is different graphs, right? So we see here that a wild type female is on average 21.7 grams. A female where you knocked out this actu gene actually is on average, not on average, but on median has a 17 gram body weight, right? So we can directly know that, okay, so if you knock out actu, you will lose 21. So you will lose around 4.7 grams of body weight when you are a female. When you are a male, the effect is a little bit bigger, right? And you can actually see all of the body parameter plots, right? And here you can see the measurements that they did from September 2007, where they measured the triangles. So the triangles here are the wild type females and you have wild type, the triangles are the wild type males, the circles are the wild type females. So a lot of wild types were measured, then they measured even more wild types and then they started measuring the homozygous knockouts and so here you can see when the measurements were done, right? It shows you a summary table. It shows you that there's a massive, massive influence of this gene on the phenotype and they also show you which statistics they used and you can actually get access to all of the data. So you can just download all of the body weight data done on the males and on the females. And this is a massive, massive resource. So if you wanna know anything about your phenotype of interest and of course this database since it is mouse, if you're interested in flowering time, of course you're not going to, it's not going to work, right? Because mice don't flower or not really. But any phenotype that you might be interested in humans, they will have some genes. So if you ever have to write a master's thesis and your master's thesis is about obesity, then this database can help you give you a list of genes which might be involved in obesity and not just that, but it generally also provides the data to you so you can redo the analysis yourself or you could look at other things, right? So if we just go back to the act two gene, right? So not just the body weight. And then we see here the summary of the gene and of course the gene does not only control body weight, right? It has a significant influence. So here with the little phenotype summary using the nice pictograms, you can see which ones are significantly affected, which were not significantly affected and which were not tested yet. So if we just hover over it, we can see that the skeleton, if you knock out act two is significantly affected. But for example, the limbs, so the digits or the tail, it doesn't have a significant influence on that. It's a perfect database to get a little bit of an overview of what's going on with my phenotype or what's going on with my favorite gene, right? So my favorite gene or not so much favorite gene because like is actually BBS7. So we can actually search for BBS7. So BBS7 is one of the genes that we study a lot because one of our mouse models has very likely a kind of mutation in BBS7. We work on a mouse model which is called the Berlin Fat Mouse and the Berlin Fat Mouse, like the name already says is fat. It's really fat compared to a normal mouse. I will probably show you guys a couple of pictures about that as well in later presentations. But then BBS7 is the gene that when we did QTL mapping, so we did show the Fat Mouse. Yeah, and then I have to look at a picture of the Fat Mouse. Let me get you guys a picture. Since it's three, I will actually stop the recording. So for the people watching it on Moodle, I will see you in part three of the lecture and you're not gonna see the Berlin Fat Mouse. So you will have to wait until the next lecture. So I'm gonna stop the recording.