[SOUND] [MUSIC] [MUSIC] Before we will go into genome wide association studies, or GWAS for short, let us consider a fate of individual mutant allele. So to start with, any mutation in human population occurs in a single copy on a single specific ancestral chromosome. Because of drift, because of random genetic events, this mutation may spread in the population. However, it's not only the mutation, but also genetic context around this mutation which spreads. And originally, in the first few generations, there are rather big stretches of DNA which co-segregate together with the mutation, while in the number of generation recombination is breaking down the structure. And the region shared by mutations identical by descent becomes shorter and shorter. Here on the picture, I have depicted this fact by painting the regions around the mutation with red. Of course, in reality, we are not really painting the chromosomes in red. However, we can trace the inheritance of specific stretches of DNA using genetic markers. Something about which Michelle was talking in some of the previous lectures. Let's now think what happens if this mutation leads to a Mendelian disease. So with few exceptions, Mendelian disease are rare and the alleles which lead to this disease have very high penetrance. At the same time, there is as a rule, absence or very low rate of phenocopies. That is, if we see the mutation, almost for certain it will lead to the disease. While if it's a disease, most likely we will see a mutation in specific gene. This means that we can study this disease via collecting material through ascertainment through proband. So we can go to a population and sample just one case with a disease. However, given this disease is controlled by rare mutation with high penetrance, this means that it's very likely that the relatives of this person are going to also be carriers of this mutation. And through proband we are going to ascertain a whole pedigree, where the number of cases is going to be enriched compared to the populational very rare rate. So what this means, that we can study this disease in relatively small families. Well first of all, close relatives are going to share very big stretches of DNA. If you take any specific chromosomes in the example presented on the screen, there are going to be only few recombination events. This means that only few hundreds of markers is enough to paint our chromosomes and to define the combination events and consequently to map the disease locus in this pedigree. Although the resolution of this method will be relatively low, we can go and ascertain different families. And through overlapping the signals for linkage in these families families we are going to fine map the region where the affected local lies. So the technology was already developed to do several hundreds of markers in 80s. And this defined very large progress in the mapping of loci of Mendelian diseases during last decades. However, now let's turn to complex traits or polygenic traits. And here the model of control is different. So we've been talking about mutation which is almost necessary and sufficient for the disease. When we talk about polygenic disease, any particular mutation is neither necessary nor sufficient to determine disease. We are talking about disease or complex phenotypes, quantitative traits, for example example, which are defined by joint effect of multiple genetic and environmental factors and their interplay. So, if we can see each of these genetic variants, we will find that they have a potentially very low, or maybe moderate effect on the phenotype. So what does it mean? This means, first of all, that if you're going to try to extend the design, which we've been using to study monogenic disease, into complex traits, we are going to fail miserably. Because if you try to ascertain a pedigree via proband with the disease, you're going to find that your proband is most likely because of one reason, while the diseased in this family are going to be because of some other reason. So this gives you effectively no power to study complex disease via linkage. But what we can do? So the key is of course the power, so you need very big sample sizes. However, the problem will be that if you sample very big samples they're going to be connected together in a huge pedigree. And this means that the region surrounding any particular mutation or a variant affecting your trait or disease of interest is going to be very small. This means, you need a huge number of markers to study this complex traits. By the end of 90s, it was theoretically figured out how many markers you would need to study disease and complex traits in huge pedigrees or huge populational base samples. And it appears that the number is about few hundreds of thousands. However, in the end of 90s, the technology didn't allow for a genotyping of hundreds of thousands markers and the genetics of complex human disease had to wait until new technologies appeared.