-
New Forum Software Installed (Still a few issues but reopening the forum) More details.
UK Genome Wide Association Study (GWAS) project - draft website goes live, feedback sought on recruitment plan, and updates
Andy
Retired committee member
Saz94
Senior Member (Voting Rights)
With well-defined inclusion criteria (see Andy's comment above mine), the size of the sample does not affect the likelihood of "mish mash" getting in. But, having such a large sample increases the statistical power, which means that the study's findings are less likely to be affected by any "mish mash" that do get into the sample.
I just find it hard to believe there are 20,000 people in UK with real ME. But that's only my humble opinion. I know here in Ireland figures are totally inflated.
Barry
Senior Member (Voting Rights)
Why do you think that?
We don’t have hard numbers such as exist for a well-defined and diagnosed disease like, ohhh, Type I diabetes. We do have a range of estimates, using varying definitions of ME/CFS/CFIDS.
Population of the UK is about 67 million. What do you believe the incidence rate of ME (‘real ME’) is?
Andy
Retired committee member
Application goes in tomorrow, we expect to hear the decision in March.
Kitty
Senior Member (Voting Rights)
Gosh, that's quicker than I expected! Excellent – will keep everything crossed.
Robert 1973
Senior Member (Voting Rights)
This is really great:
Many thanks to everyone who has helped to get it to this stage.
Just one note of cation: Despite the success in getting so many people to sign up in such a short time frame, my suspicion is that there may be a ceiling of around 10–15 thousand volunteers that could be recruited through the sort of channels that have been used thus far. That would be in keeping with the number of people who have signed some of the more successful UK online ME petitions.
Given that some of the people who sign up will not fulfil the criteria for inclusion in the study, that would still mean that a significant number of patients will need to be recruited through other channels – ie mainstream media. I also think that will help to ensure a less biased selection of patients. My understanding is that this is the intention, so hopefully that will not be a problem if it gets the funding. Fingers crossed.
Many thanks to everyone who has helped to get it to this stage.
Just one note of cation: Despite the success in getting so many people to sign up in such a short time frame, my suspicion is that there may be a ceiling of around 10–15 thousand volunteers that could be recruited through the sort of channels that have been used thus far. That would be in keeping with the number of people who have signed some of the more successful UK online ME petitions.
Given that some of the people who sign up will not fulfil the criteria for inclusion in the study, that would still mean that a significant number of patients will need to be recruited through other channels – ie mainstream media. I also think that will help to ensure a less biased selection of patients. My understanding is that this is the intention, so hopefully that will not be a problem if it gets the funding. Fingers crossed.
Andy
Retired committee member
Yes, very much so. While we are delighted with the level of support shown so far, we would like to err on the side of caution with regard to recruitment, so we're not assuming that the level of response so far means it will be easy to get to our target of 20k participants.
Kitty
Senior Member (Voting Rights)
It's excellent news, isn't it – very encouraging.
This is true, but so far no money has been available to spend on publicity, and I don't think GP practices or hospital clinics have been approached yet (it's too early to get them on board). I think if the team can utilise these channels, as well as social media and word of mouth, the reach could be much greater.
This is true, but so far no money has been available to spend on publicity, and I don't think GP practices or hospital clinics have been approached yet (it's too early to get them on board). I think if the team can utilise these channels, as well as social media and word of mouth, the reach could be much greater.
ukxmrv
Senior Member (Voting Rights)
Edinburgh Uni has another genetic study which I have provided a saliva sample for. The written material sent to me said it should take 2-5 minutes but it took me 12 minutes to eventually get up to the line on the tube (and that was with my other half coming in with a plate of yummy smelling food)
The instructions said that one should not eat, drink , smoke or chew gum for 30 mins before giving the sample but maybe this can be checked when they move into the ME group.
Just as an aside there was a questionnaire with the current study and under health conditions they did ask if one had ME/CFS but I can't remember how it was written.
Andy
Retired committee member
To expand on this, it has been indicated to us that we should receive the decision by the end of March. Just thought it would be worthwhile updating that to manage expectations now we are actually in March.
Andy
Retired committee member
Confirmation of additions that were made to the application at the last moment (this will be added to the website in due course).
Additions made to the application for funding.
In the final weeks before the application for funding was submitted, we were able to identify a number of areas where we could make some savings.
Sadly these savings weren’t sufficient to re-implement our original plan of an expansion of the UK ME/CFS Biobank, but they did allow us to add the following items:
100 whole genome sequences
Whole genome sequencing (WGS) will allow us to assess the accuracy of our GWAS genotypes and would be a pilot for a future – much larger – WGS study. We can’t make reliable claims about genetic variants causing ME/CFS with only 100 genome sequences. There is a very small chance that ME/CFS-causing genetic variants are commonly present in one or two genes and if so we would observe this, but this is unlikely. The ME Biomedical Partnership study was always going to bank a DNA sample for future WGS and this pilot study will demonstrate to funders that this future WGS study is feasible. Eventually, we hope that all DNA samples, for which consent was given, will be subjected to WGS.
60 Smartwatches
We will use these to measure wearers’ activity and skin temperature continuously over the period of one week. Once the smartwatch is returned and its data downloaded, then we can objectively compare any participant’s measured activity to anyone else’s, including individuals from the UK Biobank project. At no cost, we’ve recruited an expert (Dr Thanasis Tsanas) to help us with the data analysis.
Further development of the LSHTM CureME algorithm
We will also further develop the algorithm for identifying cases of ME/CFS, according to the CCC and IOM criteria, using electronic health records. This research will be based on clinical parameters which are available for both UK ME/CFS Biobank and UK Biobank participants, and will enable us to re-classify ME/CFS cases without a formal diagnosis according to those clinical criteria, with a good degree of certainty.
Patients with potential ME/CFS and no exclusionary conditions will be placed into four groups – where Group 1 has the strongest indication of ME/CFS in line with clinical parameters and Group 4 has the weakest indication, i.e. having been misclassified, or having another condition such as general tiredness. A random sample of 125 patients in each group will be referred to an online platform for completing a questionnaire to validate their diagnosis. We will compare features of patients with a valid diagnosis to those without a diagnosis. This specific algorithm would be a valuable output that can be directly applied to the GWAS study, as well as to clinical practice and epidemiological research using big data.
Additions made to the application for funding.
In the final weeks before the application for funding was submitted, we were able to identify a number of areas where we could make some savings.
Sadly these savings weren’t sufficient to re-implement our original plan of an expansion of the UK ME/CFS Biobank, but they did allow us to add the following items:
100 whole genome sequences
Whole genome sequencing (WGS) will allow us to assess the accuracy of our GWAS genotypes and would be a pilot for a future – much larger – WGS study. We can’t make reliable claims about genetic variants causing ME/CFS with only 100 genome sequences. There is a very small chance that ME/CFS-causing genetic variants are commonly present in one or two genes and if so we would observe this, but this is unlikely. The ME Biomedical Partnership study was always going to bank a DNA sample for future WGS and this pilot study will demonstrate to funders that this future WGS study is feasible. Eventually, we hope that all DNA samples, for which consent was given, will be subjected to WGS.
60 Smartwatches
We will use these to measure wearers’ activity and skin temperature continuously over the period of one week. Once the smartwatch is returned and its data downloaded, then we can objectively compare any participant’s measured activity to anyone else’s, including individuals from the UK Biobank project. At no cost, we’ve recruited an expert (Dr Thanasis Tsanas) to help us with the data analysis.
Further development of the LSHTM CureME algorithm
We will also further develop the algorithm for identifying cases of ME/CFS, according to the CCC and IOM criteria, using electronic health records. This research will be based on clinical parameters which are available for both UK ME/CFS Biobank and UK Biobank participants, and will enable us to re-classify ME/CFS cases without a formal diagnosis according to those clinical criteria, with a good degree of certainty.
Patients with potential ME/CFS and no exclusionary conditions will be placed into four groups – where Group 1 has the strongest indication of ME/CFS in line with clinical parameters and Group 4 has the weakest indication, i.e. having been misclassified, or having another condition such as general tiredness. A random sample of 125 patients in each group will be referred to an online platform for completing a questionnaire to validate their diagnosis. We will compare features of patients with a valid diagnosis to those without a diagnosis. This specific algorithm would be a valuable output that can be directly applied to the GWAS study, as well as to clinical practice and epidemiological research using big data.
Andy
Retired committee member
A lot of money.
To avoid me attempting a probably inaccurate explanation, I've asked if Chris Ponting can write one, or point me towards an existing one, that I can post here.
Andy
Retired committee member
One sentence 'lay persons' answer.
There are 3 billion DNA letters in a person's genome. The WGS would read all of those letters, while the GWAS would ‘only' read 1 million letters.
Far more comprehensive answer written by a geneticist.
GWAS vs WGS. These not just use different technologies but they are almost different philosophies.
GWAS uses a microarray chip whose read-out is whether a person’s genome *at a particular location* contains 0, 1 or 2 copies of a specific letter (e.g. A or C or G or T). The chip can do this read-out (“genotype”) at each of about 1 million locations. So, in other words, GWAS reads out only 1 million DNA letters. This is OK because most DNA letters that are commonly different (>0.5%) in the human population (i.e. are DNA variants) can be read out in this one chip. The great thing is that this is cheap. Another great thing is that because DNA variants that are close to one another on chromosomes are commonly inherited together (something confusingly called linkage disequilibrium) you can quite accurately guess the letters of one variant when you have read-out another close by.
Whole genome sequencing. This is what it says on the tin: sequencing of every DNA letter (3 billion of them) in the genome. (Let’s forget for now that 5-10% of the human genome is so full of repetitive sequence that it’s thrown out each time. We don’t think that these repeats matter too much.) WGS thus gives you what you find from GWAS genotyping plus all the other locations whose DNA is variable. It is expensive relative to GWAS. In fact so expensive that the analysis (below) for GWAS is not really possible for WGS.
Analysis of the data from a GWAS is a statistical test of whether a DNA variant predicts well whether someone is either a case or a control (i.e. is the variant associated with either being a case or a control). The probability statistic (p-value) from a GWAS gives an indication of whether this variant distinguishes case vs control and because millions of places in the genome are being tested at once, the multiple testing burden means that – to be truly genome-wide significant – the p-value has to be tiny, less than 5x10-8.
ETA: Slight edit to text. Removed "The “A” in GWAS is “analysis”. This" and added "of the data from a GWAS" and "(i.e. is the variant associated with either being a case or a control)".
Last edited:
Robert 1973
Senior Member (Voting Rights)
Err, doesn’t it stand for “association”?
Thanks for the excellent explanation. I sort of knew most of it, but very useful to have it spelt out so clearly and concisely.
For anyone new to this topic, I’d also recommend reading @Simon M ’s blogs on the proposed GWAS:
https://mecfsresearchreview.me/2019...-of-dna-to-help-uncover-the-causes-of-me-cfs/
https://mecfsresearchreview.me/2019/11/08/bold-plans-for-two-big-uk-biomedical-research-projects/