But this complication it seems that it won’t be able to have an application on that, because it will have thousands of genes at the end we would say “Well that’s too complicated, to have any possibility of using this information for health or for medicine”. Yeah, so it’s a bit of cup half full, cup half empty of the situation. But I think the more we know about the core genes, the more we know about ways to approach complex traits. So, there’s a new paper that came out about three months ago, so in June, I think, of 2017 from Jonathan Pritchard, talking about something called Omnigenics.
And what he points out now… it’s been known that there are about 2 or 3% of all the SNP variations in the genome, probably influence any given trait. So, that’s a hundred thousand for height, a hundred thousand for obesity, a hundred thousand for diabetes risk, a hundred thousand for education attainment. So, an enormous amount of variation and any single variant explaining a tiny fraction isn’t useful.
But he also argues that the reason is that way is because there’s a core set of genes, maybe a hundred or two hundred, that there are operating in a particular set of cells that are critical for that disease, and if that genes involved in say signal tranduction from the immunoglobulin receptor then clearly it’s critical for the biology. But anything else in that cell will automatically impact upon that little activity, and that’s the background, the very, very small amount. So that background gives us the modifiers, it gives us a sort of a standard, against which all of the more highly penetrant and environmental effects operate.
So you can be protective because of your background, or you can be at high risk because of your background, and then it takes something else major, to really push you over an edge. It’s a disease. It’s the new idea.
But you don’t think that being things so complicated will make the intervention much more difficult? You know, I think it’s also important to realize that the intervention doesn’t even have to target variable things. So the orangutans behind us, all have 5 fingers. You know an orangutan with all five of this long fingers doing what they can, but of course that’s what make them who they are, with that variability. It’s the same in disease. A lot of processes don’t have to be variable. One reason that I do more than just genetics is because it gives an insight into things that aren’t captured by variation.
So in Crohn’s disease, for example, which is one of the diseases that I study, we’ve learned from transcriptomics that the extracellular matrix and the mitochondrial function are absolutely critical, but if you do GWAS, they don’t get hits, they’re not implicated. So it’s a whole integrative system that’s changing, in any of these diseases, and GWAS is just a light on that. But still, the pathology is very specific, what’s going on is much bigger than just the individual variance. So in the end, in the cells, in the cell biology, it’s just as targetable and analyzable and we can make interventions despite the complexity of the underlying genetics.
That’s the point you have introduced, it’s a very interesting point in using a, you have been pioneer on that, using transcriptomics, that’s the genes that are really working in a given cell, in a given tissue. This information is independent of the information in the genome? So the information of the transcriptomic profile and the genes, the GWAS, are independent? I think “independent” it’s a too stronger a word. There are four reasons, I guess, 3 or 4 reasons why I like to study the transcriptome, and not just the genome. So one is the one I just gave you, that actually it’s more than a variable things that matter.
So another one is we want to open up that black box between the genotype and the phenotype. And if the genotype, we want to know what the genotypes are doing at the molecular level, okay? So if you want to understand why a Ferrari is different from a Ford, just having the parts list isn’t going to tell you. You will have to understand how those different parts of the Ferrari make it faster and stronger. Okay? And so that takes studying downstream. So we can do that with transcriptomics. Another reason is because it turns out that about, the majority, maybe as many as 90% of this GWAS variance are regulating gene expression.
So it makes sense then to study how they’re regulating gene expression, study the gene expression absolutely directly. We can do that with the same sorts of tools now; we can study all 20.000 genes at once and see what they’re doing in a very coordinated manner, and it gives us new insights. And I guess the other is, I just think it’s more fun, and more exciting, actually, get to study the transcriptome on the genome.