 So this is a joint work with Leo Kirsch who's sitting in the audience wave up there if you want to speak with him later So our lab works on trying to characterize and model patterns of gene expression in the brain if we heard a few talks about this topic And we're going to hear more tomorrow Since this is a broad audience Let me let me stress that the the transcript to them the set of genes and that are expressed and Their level of expression changes from one cell type to the other from one tissue to the other from one brain region And also as we showed this morning From one age to another But the problem is that we really don't understand the principles governing these changes We don't know what determines if a gene is going to be expressed and when and where and how So we were thinking that development of the brain should play an important role Governing these principles. So the specific question that we ask here was What patterns of gene expression in the adult human brain actually? depend on the development process So let me put when I say development let me put that in the context of what we're interested here the neural tissue start from what is known as the neural plate then it falls to a neural tube and Here color-coded you see several segments that later turn into brain structure the forebrain the midbrain hindbrain and Later on the the forebrain which turns into the cortex Expands a lot it covers the midbrain Creating two hemispheres and then it expands even more creating this wiggle wrinkled structure that you all know as the adult brain So what we wanted to look at is these expression patterns in the adult brain and we use the data set provided by the Allen Institute This says data set of postmortem human brains from six donors with hundreds of tissue sampled from each brain and for each Tissue they had the microwave. So they have the full expression profile for all genes. So we imagine you have Many many points here with a high resolution So to handle all these locations We treated all these brain regions through this ontology that the Allen Institute provided So this ontology has hundreds of regions and the important thing is that it's roughly corresponds or it is it agrees with the Development process for instance on the left here. You see Frontal regions mid regions and high and brain regions for to illustrate I'm coloring here the regions with the same color. So you see the front area is here so When we come to analyze this data The first thing you could do is to look at what regions have similar expression profile And one way to do that is just to take all regions and cluster them based on their expression similarity And when you do that you get that this hierarchical structure from the clustering Actually corresponds and closely related to the ontology structure For instance, you have front areas here Hind areas here But but of course this is clustering that is based on the full genome And it is possible that just a small number of genes drive these similarity between regions So what we actually wanted to do to know is what regions what sorry? What genes drive this similarity between brain regions? so to do that we went and defined an index that was per single gene and Here's here's what we basically did. This is an example for one gene. You're a D1 So basically you take a pair of tissues and you measure two types of distances One is just how similar their expression is Kind of expression distance and the other is the distance between these tissues on the ontology tree and you Repeat that for all pairs of tissues and you get a joint distribution of the expression distance and the ontology distance And for this gene you see a strong correspondence Which basically means that if you take two areas that are close in the ontology They they have similar expression for this specific gene and you can quantify that Using some sperman correlation. So this is a measure for an individual gene now we repeated that over many many genes over all genes and what I'm showing you here is this distribution of this score which we call the brain region ontology agreement score and So each curve here correspond to another donor and it is compared with the shuffle data so what we find here is that in practice almost all genes like 92% of the genes have a Significant agreement with a brain ontology. So This is interesting because it means that when you look at the expression level of a gene It actually reflects the embryonic origin of the region where you measure the gene Not only the function of the gene. So you should take that into account Now we then repeated that in another data set that was discussed this morning data set provided by the system group the brain spend date Using adult brains only So here I'm showing here again the distribution of this bro agreement score for the two data sets Each point here is a gene and you see there is actually a strong agreement between the two data sets Even though these are different regions and different donors, etc So what can when you tell about these genes? I'll just do that very shortly We can look at genes that are expressed specifically in some cell types So you see here that neurons genes expressing neurons have much higher brain Agreement with the ontology while asteroid site has less if you only focus on cortical neurons. This is actually reverse. So this is interesting two more interesting stories is if you look at Developmental genes like axon guidance they also and this is again in the adult brain They still retain Differential expression across regions even though their main function is during development. So it's possible to have other Functions that we need to study Finally housekeeping genes are a genes that are Involving fundamental functions like basic metabolism and again, it's interesting that actually the brain uses different genes in different areas So just a quick Conclusion what we find here is that the aerial expression patterns actually are in agreement close agreement with ontology that agrees with the development And also it's interesting that many of these develop families of development genes May have some other functions in the adult brain that we should look at So so far I told you about Spatial patterns of expression. I want to spend two minutes on temporal patterns of expression. I We focused on the period of adolescence. I have two teenager kids at home So you can met some of you can sympathize why this was personally interesting So in general the lessons is a period that has Teenagers go under massive Behavioral cognitive changes and also there's a much higher prevalence of psycho pathologies Including mood disorders but depression and aggressive impulsive behavior and it's of course very interesting We actually don't know that what could be the molecular mechanisms underlined these these strong changes So we focused on two specific systems the serotonin system and the dopamine system since these are known to be tightly Related to disorders like depression Drug targets, that's it right ssri And what we did was basically to look at all genes in these two systems equal in pre-synaptic post-synaptic all these genes signal transduction and Basically what we did I'm skipping the detail is to measure for each gene How strongly their expression changes from childhood to adolescence? so what I'm showing here is a Heat map where each rose corresponds to a different gene and column to a different brain area this is the brain span data again and The the color here correspond to the magnitude of the change between a childhood and adolescence So what do we find? First we hardly see anything in the dopamine system, but we do find two genes that have strong transition Between childhood and adolescence in the in the serotonin system Here's how the measurements actually look and I don't I don't think you could see but there's a there's a period here that we Considered as the adolescence. So you see this mark change now these genes are actually very interesting They're not just any genes these are auto receptors. So sorry these genes code for auto receptors These are receptors that are located on the pre-synaptic side and allow the pre-synaptic cell to monitor How much serotonin is secreted into the cleft? So this is kind of a feedback Control loop and it's and it's pretty interesting that this control mechanism is one that changes abruptly in in adolescence So let me conclude we found that special expression patterns in the adult human brain Are actually tightly related to do the the brain region ontology? Corresponding to development and this happens for the vast majority of gene more than 90% of the genes And I think this is important because it means that when you try to interpret a gene expression pattern in an adult brain You just have to take into account this baseline that is not just uniform across regions Second we found these two specific serotonin auto receptors that exhibit this a huge expression transition during from childhood adolescence and this is important because it puts a focus on on a possible Mechanism involving these changes, of course, we don't know if this is involving depression and we're doing some follow-up work So thank you very much So there are a lot of microarrays taken in Let's say like depressed people and control groups. There are no time forces and also the Ages at which you get these these brains is very limited. So So we are planning to do experiment in mice One thing I didn't say is that this transition when once we found is a these results We went and looked at other species. It actually happens in monkeys, but we don't see for these specific genes in mice We don't see that so this could be a primate result human development one is lepska and Weinberger and The other one is ours Geo data sets. Okay. All development for frontal cortex, right, right? Yeah So so so for depressed. No, I think the question was for the present development Right. There is there is also one for depression. I can I can