Gene expression meta analysis study, and a blind trial, proposed by Professor Kellam

Dolphin

Senior Member
Messages
17,567
Wait, where is the discussion of a blind trial? Is this going to be based on what was predicted by the meta analysis, or have I filled in too many blanks?

From their latest newsletter:
Moving towards a diagnostic test:

We have come a long way since Dr Rob Powell started the first gene study Altered gene expression in CFS , firstly under Professor Jeff Almond at the University of Reading and then under Professor Stephen Holgate at Southampton University. Until then research worldwide had studies different aspects of the illness but as a result of this study, and the information received from other disease studied funded by the Foundation the Research Committee decide to study the basis of the disease. Scientists were studying the involvement of genes in many illnesses and the Research Committee decided that a study should take place to compare the genes of CFS/ME patients with those of normal healthy people.. Dr Jonathon Kerr, who has already carried out some research funded by the Foundation, was invited to submit a grant application. He planned an ambitious project using scientists from laboratories in three British Medical Schools and one in the USA. This project needed courage because in gene research, samples taken from lesions are studied but there are no lesions in CFS/ME so the scientists decided to use white blood cells. This had never been attempted before in any illness so there were several nail biting years while we waited to see if ythis could be successful. When Dr Kerr announced that he had found 88 genes in CFS/ME patients that were abnormal but remained normal in the healthy group, CFS/ME research was changed worldwide. Dr Kerr took his research further and published a paper in the Journal of Infectious Diseases; a journal which is known for its very high standards, suggesting that the 88 gene profiled could be divided into 7 subtypes.

The CFS Research Foundation has always insisted on the highest standards of research and the ability to reproduce the results of ant study is always paramount. Professor Paul Kellam, Dr Jonathon Kerr, Dr Tim Harrison and Dr Dan Frampton carried out a blinded study to determine whether the gene expression profile was able to discriminate between patient and controls.

Unfortunately this was not the case. When the study was completed, Dr Ker decided to return to clinical work but the gene research is to continue building on Dr Kerrs work.

After the pioneering work of Dr Kerr and colleagues was published other research groups throughout the worlds have carried out similar studies, but there have been some differences in the results. This was to be expected as there was some variation in the technical features of the scientific methods used, in the classification of patients and where the study was conducted. But in all studies results showed clear differences between patients and controls.

Clearly this needed further study so Professor Paul Kellam, Virus Genomics Team Leader at the Welcome Trust Sanger Institute, Cambridge together with Dr Tim Harrison designed a 2 part study, Meta analysis of gene expression data to discover diagnostic and mechanistic signature in the peripheral blood of CFS patients. They submitted this study as a grant application to the CFS Research Foundationm and the Foundation agreed to fund it for 2 years.

Because the different laboratories in different coutries have used a similar approach in their research it is possible to analyse their results in combination using powerful computing and a method called meta analysis. Using this advanced computing the results of these studies will be analysed together with those derived from Dr Kerrs work. The advantage of this approach is that analysing the studies in combination is much more powerful than any of the studies considered individually and the team expect that the original group of 88 genes will be modified to generate a more formidable set. In addition, if small effects that dont appear to be significant in one study can be reported across several studies, they will have a better idea of which genes are involved in CFS/ME. Without combining these studies it would be impossible to identify such subtle effects.

Currently this group is chasing gene expression data from 8 CFS/ME studies concluded across the world, from the Uk to the USA, from Japan to Sweden. The detective work of finding the raw data is only half the battle, the second is ensuring that the studies are comparable, the data is compatible and the CFS/ME classifications behind each study are reliable.

These scientists are aiming to combine these existing studies from around the world, analyse them and identify such genes that are changed in this way across each of the studies what are called CFS marker genes. They will then test their predicted marker genes in a separate study using more blood samples from both CFS/ME patients and healthy volunteers. If gene expression patterns can be reliably produced they will have successfully identified a robust set of CFS marker genes that can be used as a diagnosis.

One issue they need to consider is that there exists several subtypes of CFS/ME some with more severe symptoms than others and these might be expected to have different patterns of gene expression. By combining data from several studies it may be possible to identify the underlying subtypes and noting key differences in the gene expression patterns which might have been missed on an individual study or attributed to samples variation. The outcome of this will be a modified stet of genes which can be used potentially to discriminate between patients and controls.

They will then proceed to the second part of this study when they will test the findings of the first part for accuracy. They will recruit a new set of patients and controls when the expression of a new set of genes will be tested in a blinded study. The samples will be placed in tubes, each one coded then frozen. One of the scientists will keep the code and no-one else will know it until the time set for unblinding. It will then be possible to test the accuracy of teh first part of this study.

If this 2 part study produces gene expression patterns which can be reliably reproduced Professor Kellam and his team will have successfully identified a robust set of CFS marker genes which can be used as a diagnosis.

This is likely to have a profound effect on CFS/ME research. The problem we have been battling with at the Foundation has been lack of research applications of a sufficiently high standard. We have to acknowledge that the scientists are hesitant to undertake research into a disease for which there is no diagnosis, but interest in the possibilities of research into CFS/ME has been growing. Following the publication of the XMRV paper groups throughout the world tried to reproduce this study. One journalist said that scientists were leaping out of bed in the mornings to study the XMRV. Sadly no other groups were able to reproduce this study and subsequent studies have found the problem to be contamination. If scientists can feel confidence in a diagnosis then surely they will come to us with new ideas for research projects. So much hangs on this study which is being undertaken by Professor Kellam and his team. A diagnosis would change the way CFS/ME is regarded by doctors and scientists alike.

It would be wrong to easily assume that this ambitious study will yield the answers we wish for. There are two years of hard work ahead, but the study has been carefully planned and has enjoyed the criticism and help of eminent scientist s specialising in this field. It has the greatest chance of success and we shall all be awaiting the results with hope.

The research funded by the foundation has made enormous strides since 1998 and we must be grateful to the scientists who have taken us forward with such speed that we can hope that we may soon have a diagnostic test. All this would not have been possible without the outstanding generosity of our supporters who have not only funded our research but have given us such encouragement throughout the years. We must now brace ourselves for the demands of this new and exiting study. We have achieved so much that we must confidently expect that we shall successfully take forward our research. We are on the point of finding a diagnostic test and that would lead us to new therapies and then a cure. A few years ago that would have been only a dream but now it is within our sights.
Anne Faulkner. Hon Director
 

Dolphin

Senior Member
Messages
17,567
Dolphin: Really good point about not including CDC-empiric defined patients - could you let me know the relevant CDC paper(s) and I'll try to write?
Quick response: the CDC have done gene expression research on a few groups incl. one post-exercise study.

The empiric-criteria study is:
Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, Papanicolaou DA, Unger ER, Vernon SD, Heim C. Chronic fatigue syndrome--a clinically empirical approach to its definition and study. BMC Med. 2005 Dec 15;3:19.
so one is talking about the data that came out after that paper, which is spread out over lots of papers e.g. Pharmacogenomics papers in 2006 (Wichita 2-day study cohort) and the Georgia cohort (gave a prevalence of 2.54%).
 

eric_s

Senior Member
Messages
1,925
Location
Switzerland/Spain (Valencia)
Following the publication of the XMRV paper groups throughout the world tried to reproduce this study. One journalist said that scientists were leaping out of bed in the mornings to study the XMRV. Sadly no other groups were able to reproduce this study and subsequent studies have found the problem to be contamination.
I would say this is wrong. What about the Hanson group? Besides that group, XMRV was also found in other labs, but you could say they didn't reproduce Lombardi et al. because the number of samples was too small or for other reasons.
And yes, other studies have concluded that the problem is contamination, but that does of course not mean they are correct. This remains to be seen.
We are on the point of finding a diagnostic test and that would lead us to new therapies and then a cure.
That would of course be great, but there are also other approaches for diagnostic tests, like cytokines, immune system markers, spinal fluid, XMRV and probably others.
I also think even if they are able to find a pattern of gene expression it might still be quite a long way from there to a therapy. Because probably one would have to find the cause of that change in gene expression. But either way, even if i don't like what they say about XMRV, any serious research is good.
 

oceanblue

Guest
Messages
1,383
Location
UK
The CFS Research Foundation's website says they have funded a lot of Dr Kerr's research. Their website lists a blinded study that Dr Kerr carried out, using his gene expression findings, and apparently that study failed. I don't think I've ever heard about that failed study before now, as I would have been very interested in the results, and I don't remember hearing about it.

Sep 08 - Jul 10
Blinded class prediction (Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME) versus Normal) using gene expression data and an artificial neural network (ANN)
Grant = 135,514
Dr J Kerr, St. Georges University of London

http://www.cfs-research.org/projects.htm
Looks like this is it:
Assessment of a 44 Gene Classifier for the Evaluation of Chronic Fatigue Syndrome from Peripheral Blood Mononuclear Cell Gene Expression Daniel Frampton1, Jonathan Kerr2, Tim J. Harrison3, Paul Kellam1,4*

Chronic fatigue syndrome (CFS) is a clinically defined illness estimated to affect millions of people worldwide causing significant morbidity and an annual cost of billions of dollars. Currently there are no laboratory-based diagnostic methods for CFS. However, differences in gene expression profiles between CFS patients and healthy persons have been reported in the literature. Using mRNA relative quantities for 44 previously identified reporter genes taken from a large dataset comprising both CFS patients and healthy volunteers, we derived a gene profile scoring metric to accurately classify CFS and healthy samples. This metric out-performed any of the reporter genes used individually as a classifier of CFS.

To determine whether the reporter genes were robust across populations, we applied this metric to classify a separate blind dataset of mRNA relative quantities from a new population of CFS patients and healthy persons with limited success. Although the metric was able to successfully classify roughly two-thirds of both CFS and healthy samples correctly, the level of misclassification was high. We conclude many of the previously identified reporter genes are study-specific and thus cannot be used as a broad CFS diagnostic.
 

Enid

Senior Member
Messages
3,309
Location
UK
I don't know anything about the science but all things considered this has to be welcome research (especially in the negative climate created by the psych lobby in the UK)
 

Esther12

Senior Member
Messages
13,774
To determine whether the reporter genes were robust across populations, we applied this metric to classify a separate blind dataset of mRNA relative quantities from a new population of CFS patients and healthy persons with limited success. Although the metric was able to successfully classify roughly two-thirds of both CFS and healthy samples correctly, the level of misclassification was high.

tbh - that's better than I expected.

I'd be pretty comfortable betting that it's also better at distinguishing between CFS patients and controls than any mental health questionnaire.
 

oceanblue

Guest
Messages
1,383
Location
UK
tbh - that's better than I expected.

I'd be pretty comfortable betting that it's also better at distinguishing between CFS patients and controls than any mental health questionnaire.
True enough. I think Lenny Jason has a couple of Qs on PEM that are >90% accurate on patients and controls: that's the benchmark.
 

Esther12

Senior Member
Messages
13,774
True enough. I think Lenny Jason has a couple of Qs on PEM that are >90% accurate on patients and controls: that's the benchmark.

Well - you could always have a questionnaire asking 'Do you have CFS?' - but that's kind of cheating. I meant ones measuring illness beliefs, personality, coping techniques, views on mental health, etc.
 

Bob

Senior Member
Messages
16,455
Location
England (south coast)
CFS Research Foundation - Full newsletter - corrected version

The CFS Research Foundation have sent out a new newsletter saying that the previous one had been re-written and distributed without their permission, and had mistakes in it.

So they have sent out an official corrected version.
I've attached it as a PDF file.
 

Attachments

  • CFS Research Foundation - NEWSLETTER FINAL.pdf
    127.1 KB · Views: 23
Back