The science behind genetics and the genetics of ME/CFS:
Chris: Thank you, Shona, and it’s a delight to have you as part of the team. So what we’re going to do now is I’m just going to talk you through some of the science behind genetics and about the genetics of ME, and then we’re going to go into a Q and A session. We’ll be fed some questions live and also we have some questions that people have sent in ahead of time. I apologize ahead of time, there might be some background noise in the house, there are people who are working nearby and I apologize if that happens.
So what I’m going to do is share my screen and hopefully now you’ll see some slides and see that this is a project that is run for us out of the MRC Human Genetics Unit. It’s funded by the Medical Research Council and the National Institutes of Health research.
The important thing is to say at the beginning is that we know that ME/CFS is in part genetic, and that’s because of previous research such as this one done over ten years ago, showing that the relative risk of someone being diagnosed here with CFS is increased (and this is in the red box) if they have a first, 2nd or third degree relative.
So you have about two-fold greater likelihood of being diagnosed if you have a first, 2nd or third degree relative who’s also been diagnosed with CFS. And there is other information around which point in the same direction. So what do I mean? And I said this earlier, what do I mean by the fact that we have letters that are different in our genomes? So what I’m showing here is two copies of a genome and you have two copies because you gain one from your mother and one from your father. So if you compare those two copies side by side, you will see that essentially they will be identical except in about one in 1000 places where there will be a difference. And this difference we’re going to call “SNIP”, which stands for single nucleotide polymorphism, but I’m also going to call it a letter change. And here the letter is changed from a C or G to an A or a T. And I’m saying C or G because they are paired in a double Helix.
Anyway, the important thing is that there are these changes and people have looked at these changes and have asked whether the changes predict whether someone is more or less likely to be diagnosed with ME or CFS or ME/CFS. So here is one such study, which has said that the letters in a gene called the glucocorticoid receptor gene and has said that there are letters that predict whether someone has chronic fatigue syndrome.
That such study has been done in a whole variety of different genes and different parts of the genome and what we have more recently done, and we’ve published this in a review with Joshua Dibble and Simon McGrath, is to say let’s look at those predictions and ask whether in a very large data set in the UK Biobank, whether we are able to see those predictions repeated. Unfortunately, no, they’re not in general repeated. If they were, then these bars on the left here in blue would be very high and these bars on the right here would be very low. That is not the case. So, unfortunately, we don’t see any replication of gene by gene comparisons of letters.
So instead, we look at the whole genome.
We look at the genome wide associations and this is what we’re wanting to do in DecodeME, is to have a genome wide approach, a genome wide association study, a GWAS. What this does is look at the DNA in chromosomes, and these are the sock-like structures at the bottom here, and at every point in which there are changes that are common in people, we ask whether those common changes are more frequent among people with ME versus others who don’t have ME If this is so that there are more people with this letter in cases with people with ME than in controls, and this is statistically significantly so, then they predict whether someone is a case, someone with ME or control. What’s important to say at this point is that these letter changes do not categorize people. They don’t 100% separate between people with ME and controls.
So here’s a new article that’s just come out by colleagues in Norway who have looked at people with ME, ME/CFS in three different cohorts; the Norwegian cohort, a Danish replication cohort, and a data set from the UK Biobank, which is what I was just referring to.
It’s a very interesting study and a very well-done study. What they conclude, and I’m just cutting to the chase here, they say that we did not find any ME/CFS risk loci displaying genome-wide significance.
And pictorially, this is what they see – they see these dots across the X axis here, where all of these are different chromosomes, each of the dots is a place where these letters commonly are different among our genomes. If we were to have one of these dots higher than this red line, higher than this point here, then that would indicate significance at the genome wide level. There are none – there are no such points, so they said we did not find any associations at the genome-wide significance level.
They did indicate that there might be some association to a particular gene called TPPP. They said that there was modest association, but when you look at the large data set for CFS in UK Biobank, again, there are no such dots that go above this red line.
So no strong evidence for this gene. What they then say, and I appreciate what they’ve said here, that DecodeME will enable, as it is, a larger and more powerful cohort, will enable a more definitive investigation of associations between genetic risk factors of ME So that points to our study.
So our study is questionnaire’s from home. We send a saliva collection kit to your home and we ask you to send that back via the regular post. And we do all of the DNA processing and the data analysis away from your home. So we ask you to do questionnaires either online, which is our preferred option, or if that presents difficulties, then via paper questionnaire. We ask you to do that if you are expected to pass criteria – that means that you’ve had a professional diagnosis of ME or CFS or ME/CFS, you’re 16 years of age or older and you’ve got a diagnosis of any type of ME, CFS or ME/CFS., mild, moderate or severe.
We are wanting to recruit to the study 25,000 people, 20,000 people who were diagnosed because of an onset prior to the COVID pandemic, and 5,000 who specifically have been diagnosed with ME/CFS because of their COVID-19 virus infection. And that is a newer part of our studies since last year.
So we are going to for those people who pass the criteria (and if you don’t pass the criteria, it doesn’t mean to say you don’t have ME/CFS, it does means that we just have particular criteria that have been set for us internationally and by our funders) and so we will ask for those people who do pass those criteria to give a saliva sample in a tube that goes back into an envelope, that goes back into the post and is sent to the biosample centre in Milton Keynes, who will extract the DNA.
And from there, the DNA goes over to California, and it is analysed there in exactly the same way as was done for the UK Biobank and they send us back the data. All of this is done in an anonymised way, and we have strict confidentiality agreements with all of our partners, which have been thoroughly vetted by our institution and by the ethics committee. So then we do the DNA data analysis.
We ask this question of every part of the genome, whether there is a letter that is substantially and significantly more frequent among people with ME than in others. We’re able to use the UK Biobank in general for our controls.
So that’s it, we will hopefully be able to find these dots that lie above the red line which are indicated here in these red circles.
And what does that show us? That shows us that there are genetic risk factors for being diagnosed with ME But that’s not where we stop. We will go as fast as possible to understanding the genetic contributions to ME as part of what we already know about biology. Through disease networks, how molecules work together. We’ll use every type of approach that we can. High power statistics, artificial intelligence, machine learning we’ll fold in drug databases, we’ll ask whether existing drugs are likely to reverse these symptoms observed because they are targets of the genes that have been pinpointed by our genetic study.
And then it’s not up to us, but up to pharmaceutical companies to take those target genes that we’ve been identifying and to take them forward for trials which take, unfortunately, very many years and involve many thousands of participants and can take, as it says at the bottom right here, several years to complete.
But the reason that we start with genetics is that this whole process of going to trials has a very high drop-out rate. So the number of targets that go forward drops very substantially very quickly. But if you have genetic evidence behind those targets, then that reduction of targets over time is very much slowed. So there are so many people I have to thank. I just want to do that here on this slide.
Lots of people working very, very hard indeed. So thank you very much to all of them. And please do pass on to anyone our DecodeME website and do get involved. So I’m just going to stop sharing now.
So hopefully, I’ve been helpful to you to understand a little better of what the DecodeME project is and why we’ve been looking at genetics.