Stanford Community Symposium 2018: Phair, Metabolic traps, Tryptophan trap

[RDP]

Finally, does the metabolic trap theory offer any explanation as to why there is such diversity of severity of illness among ME patients? If so, what is the explanation? Is the suggestion that illness severity could be determined by the degree of IDO 1 inhibition? If so, what would determine the degree of inhibition? Or is the assumption that illness severity may be determined by other factors once the trap has been triggered?

This question can be seen as a variant of your sudden vs. progressive onset question. Trapping more cells of a given cell type or trapping new additional cell types both could lead to increased disease severity.

Good questions. Thanks for your interest.

 

[Barry]

I dont have any technical knowledge about this, but I have wondered if the severity of ones illness could in part have to do with the number of entrapped cells, and not just the degree of entrapment at the cellular level, since the metabolic trap seems to be a cellular phenomenon rather than a systems phenomenon.

I had not appreciated the significance of this until @RDP commented on it. Although I appreciate the trap is at a cellular level, I had just thought cells would fall prey to it wholesale, rather than maybe being selectively affected. For me that is a profound insight.

I wonder if for some people their pacing-level exercise helps some of their cells to flip back out of their latched state, whereas something like GET maybe just pushes more and more cells into the trap?

I also find it intriguing that if severity is determined by % of trapped cells, then does that suggest the % of trapped cells levels out for many people, given how many people's severity levels out for longish periods of time? And if the % of trapped cells does tend to a steady state for periods of time, are they necessarily always the same cells? Or could there be a steady migration of cells into and out of the trapped state, but with the average level of trapped cells fairly static? What would it be that led to this steady state condition? Some form of self regulation? Or some natural limit condition?

In reality I appreciate that even if it turns out that disease severity is significantly influenced by % trapped cells, there will inevitably be other influencing factors also. Fascinating science, but a bugger of a disease.

My scientific and medical knowledge are close to zilch, just to be clear.

 

[Robert 1973]

Many thanks for taking the time to answer my questions @RDP. Very helpful, although it will take me some time to digest your responses.

In the meantime, one more thing that puzzles me is this:

So the pivotal feature of IDO2 SNPs is not that there are so many of them, but that a few of them are both damaging (missense/nonsense, the gene product is non-functional) and extremely common. Indeed four of these variants have allele fractions > 0.01. It's because of this that one cannot just look at the UK Biobank Manhattan plot for IDO2 even if we set aside the thorny question of correct diagnosis. We are not asking if a particular SNP is significantly more common in the ME/CFS population than in the population as a whole. We are asking if at least one copy of the IDO2 gene in each ME/CFS patient is broken. So far, the answer is yes. (n=66 and counting)

@Chris Ponting (or @Simon M) can you confirm whether or not a GWAS would be capable of revealing the type of correlation that Dr Phair describes? If not, would it be possible to analyse the existing data from the UK Biobank in a different way that would determine whether or not such a correlation exists?

 

[Chris Ponting]

Many thanks for taking the time to answer my questions @RDP. Very helpful, although it will take me some time to digest your responses.

In the meantime, one more thing that puzzles me is this: @Chris Ponting (or @Simon M) can you confirm whether or not a GWAS would be capable of revealing the type of correlation that Dr Phair describes? If not, would it be possible to analyse the existing data from the UK Biobank in a different way that would determine whether or not such a correlation exists?

I'd need to read the methods describing the approach taken before I can come to a view. But I think what was being said was that a GWAS would *not* be able to reveal this result because each DNA variant is not statistically significant when considered alone. Rather, it was being proposed that all of the IDO2 variants (SNPs) that are proposed to be deleterious when *considered together* might be significant. This requires a "burden test" (and a multiple testing correction applied that controls for the number of genes plus IDO2 that were considered individually). Are there methods and results written up (& peer reviewed)? Apologies for these questions, I lost the thread of this discussion a while back.

 

[Robert 1973]

I'd need to read the methods describing the approach taken before I can come to a view. But I think what was being said was that a GWAS would *not* be able to reveal this result because each DNA variant is not statistically significant when considered alone. Rather, it was being proposed that all of the IDO2 variants (SNPs) that are proposed to be deleterious when *considered together* might be significant. This requires a "burden test" (and a multiple testing correction applied that controls for the number of genes plus IDO2 that were considered individually). Are there methods and results written up (& peer reviewed)? Apologies for these questions, I lost the thread of this discussion a while back.

As far as I’m aware there are no papers published on @RDP ’s metabolic trap theory yet, but your summary is consistent with my understanding of what has been said. (Although I note that Dr Phair has said that he did not apply a Bonferroni correction when analysing the data.)

I would be very interested to know if the available data from the UK Biobank could be used to determine whether or not it supports Dr Phair’s theory.

More generally, I would be interested to know whether a different algorithm could be applied to data collected from any GWAS study to determine whether any significant correlations of this type exist – ie between an illness and a gene that can be broken by any one of a number of different SNPs, but where no one SNP has a strong enough correlation to reach statistical significance.

 

[Amw66]

As far as I’m aware there are no papers published on @RDP ’s metabolic trap theory yet, but your summary is consistent with my understanding of what has been said. (Although I note that Dr Phair has said that he did not apply a Bonferroni correction when analysing the data.)

I would be very interested to know if the available data from the UK Biobank could be used to determine whether or not it supports Dr Phair’s theory.

More generally, I would be interested to know whether a different algorithm could be applied to data collected from any GWAS study to determine whether any significant correlations of this type exist – ie between an illness and a gene that can be broken by any one of a number of different SNPs, but where no one SNP has a strong enough correlation to reach statistical significance.

Has something like this not been done for GWI? Broderick is known for in depth " fancy math"
Otherwise, outwith psychiatric conditions, my guess would be some of the established/ better funded research diseases - cancer/ arthritis/ alzheimers and perhaps epilepsy?

 

[Jaybee00]

@Robert 1973

@RDP mentioned on PR that he had submitted a trap paper for review, but that the journal was having difficulty finding reviewers with a background in both MECFS and system modeling, so the review is delayed a bit apparently.

 

[Barry]

having difficulty finding reviewers with a background in both MECFS and system modeling

That does not surprise me at all. I think @RDP is one of a rare breed at the moment.

 

[Amw66]

That does not surprise me at all. I think @RDP is one of a rare breed at the moment.

Systems and maths and knowledge of ME - few and far between . off top of my head - Gordon Broderick who has worked with Klimas and Craig Robertson who works with Myhill ( his premise on mathematical catastrophe theory as applied to ME is kind of like a bistability switch)

 

[Chris Ponting]

As far as I’m aware there are no papers published on @RDP ’s metabolic trap theory yet, but your summary is consistent with my understanding of what has been said. (Although I note that Dr Phair has said that he did not apply a Bonferroni correction when analysing the data.)

I would be very interested to know if the available data from the UK Biobank could be used to determine whether or not it supports Dr Phair’s theory.

More generally, I would be interested to know whether a different algorithm could be applied to data collected from any GWAS study to determine whether any significant correlations of this type exist – ie between an illness and a gene that can be broken by any one of a number of different SNPs, but where no one SNP has a strong enough correlation to reach statistical significance.

Yes, there are tests for the burden of rare variants, for example:
(1) https://www.cell.com/ajhg/fulltext/S0002-9297(08)00408-4
(2) https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1000384
These work best when most rare variants alter trait (ME/CFS risk) which is not likely to be the case. These are not straightforward because gene length varies, and because 'random' mutations are never entirely random.
Unfortunately these tests cannot be applied to GWAS data because such studies *sample* DNA variants in intervals across genes and genomes, rather than sequencing every DNA letter in every gene. The sequencing approach is much more expensive (x10-20).

 

[Andy]

Paper published detailing the theory, https://www.s4me.info/threads/the-i...ology-of-me-cfs-2019-kashi-davis-phair.10517/

 

[Jim001]

I was only able to read a fraction of the paper, so I am lacking some context (and lots of technical insight), but I did find this sentence a little disheartening. I hope he is just being extra conservative.

"While it is possible that the IDO metabolic trap lays bare the etiology of ME/CFS, the probability that this is so is small."

 

[Chris Ponting]

Yes, there are tests for the burden of rare variants, for example:
(1) https://www.cell.com/ajhg/fulltext/S0002-9297(08)00408-4
(2) https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1000384
These work best when most rare variants alter trait (ME/CFS risk) which is not likely to be the case. These are not straightforward because gene length varies, and because 'random' mutations are never entirely random.
Unfortunately these tests cannot be applied to GWAS data because such studies *sample* DNA variants in intervals across genes and genomes, rather than sequencing every DNA letter in every gene. The sequencing approach is much more expensive (x10-20).

Just read the genetics part of this paper (https://www.mdpi.com/2075-4418/9/3/82/htm) and because Kashi et al. propose that "any predisposing genetic mutation must be common" I turned once again to the huge UK Biobank study with 2017 individuals with self-reported diagnosis of ME/CFS. UK Biobank data does not support this hypothesis involving common DNA variants.

 

[Hoopoe]

What does "deemed significant" mean?

Ron Davis recently said at the Invest in ME conference that 75% of the population have IDO2 mutations, but 20 severe patients and 46 additional patients tested all had IDO2 mutations. This is very unusual.

Something is very wrong. Either their method, or this biobank data.

 

[Chris Ponting]

What does "deemed significant" mean?

Ron Davis recently said at the Invest in ME conference that 75% of the population have IDO2 mutations, but 20 severe patients and 46 additional patients tested all had IDO2 mutations. This is very unusual.

Something is very wrong. Either their method, or this biobank data.

"Deemed significant" - this is the conventional genome-wide threshold, i.e. an industry standard. See https://www.ncbi.nlm.nih.gov/pubmed/30349118 or https://www.biorxiv.org/content/10.1101/176834v2 if the former is behind a paywall.
The UK Biobank data has been acquired and analysed to very high, international standards. This data set is human genetics' "go to" data set internationally.

 

[Hoopoe]

I heard that the software used by Phair to analyse genetic data ignores IDO2 mutations because they are so common. Maybe something like this is happening with the Biobank data and it's only showing what is considered to be potentially important which might only be very rare mutations.

Another possibility is that self-reported CFS could mean people that may have had no medical knowledge where asked if they have chronic fatigue syndrome and many said yes because they don't know any better and think it means being tired a lot.

Maybe the tech they used to determine whether someone had IDO2 mutation is faulty with a 100% false positive rate.

 

[Chris Ponting]

I heard that the software used by Phair to analyse genetic data ignores IDO2 mutations because they are so common. Maybe something like this is happening with the Biobank data and it's only showing what is considered to be potentially important which might only be very rare mutations.

Another possibility is that self-reported CFS could mean people that may have had no medical knowledge where asked if they have chronic fatigue syndrome and many said yes because they don't know any better and think it means being tired a lot.

Maybe the tech they used to determine whether someone had IDO2 mutation is faulty with a 100% false positive rate.

The UK Biobank results were generated entirely blinded without prejudice with respect to genes or traits/diseases. It was designed to determine whether common DNA variants can help predict an anthropometric trait or disease (e.g. whether someone is a ME/CFS case or else a control). In this case, Kashi et al. proposed that common variants can predict ME/CFS disease. The UK Biobank is supremely well placed to address this, and was very far from providing supportive evidence.

"Self-reported CFS". The self-report was whether they have ever been diagnosed as having CFS (i.e. ME/CFS) and not whether they were self-diagnosing ME/CFS. One of the central members of the UK Biobank told me that she thinks that "self-reported" data can be more accurate than electronic health records essentially because individuals are experts on what clinicians have diagnosed them with.

 

[Hoopoe]

"Self-reported CFS". The self-report was whether they have ever been diagnosed as having CFS (i.e. ME/CFS) and not whether they were self-diagnosing ME/CFS. One of the central members of the UK Biobank told me that she thinks that "self-reported" data can be more accurate than electronic health records essentially because individuals are experts on what clinicians have diagnosed them with.

Okay.

I found what I mentioned previously:

Ron Davis, on the IDO2 mutations. "If you do the automated analysis, it doesn't show up. Simply because it's excluded by the software because it can't possibly be important because it has too many mutations".

https://youtu.be/CGWcA1xYnyY?t=2057

 

[Hoopoe]

Ron Davis, on the IDO2 mutations. "If you do the automated analysis, it doesn't show up. Simply because it's excluded by the software because it can't possibly be important because it has too many mutations".

https://youtu.be/CGWcA1xYnyY?t=2057

A bit later he says that "sequencing is about 1000 dollars per patient so they decided to do it differently [...] so he worked out a way to sequence this gene for 10 dollars."

If this new way is flawed then that could explain why they're finding IDO2 mutations in all patients.

 

[rvallee]

I was only able to read a fraction of the paper, so I am lacking some context (and lots of technical insight), but I did find this sentence a little disheartening. I hope he is just being extra conservative.

"While it is possible that the IDO metabolic trap lays bare the etiology of ME/CFS, the probability that this is so is small."

I think that may be simply to mean that it could be an explanation, not necessarily the whole, cause and mechanism. That's fine. There could be a downstream process that causes this, the true cause, with the trap being a consequence that can still be reversed.

Good researchers are careful not to make claims without significant evidence. Bad ones just make stuff up. Different strokes.