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

Great questions.
(1) Until the experiment is done we will not know for sure. But because natural selection is good at weeding out "bad" mutations over 100s/1000s of generations, we must have a strong expectation that the combined variants (cancelling each other) produce a "good", functioning protein. The alternative hypothesis - that large numbers of people carry one version that produces a nonfunctioning protein - must be much less likely.
(2) "how hard is it to know what questions to ask in order to do so?" Hard - because you need to guess ahead of time what questions are most relevant without knowing what the genetics will uncover. In the project that is being planned we will use questionnaires developed and in common use by the LSHTM CureME Biobank. But this is why recruiting a cohort who are happy to be recontacted with new questions every year or so is a good way to go.

 

Hi, please bear with me as I have severe M.E and struggle a lot cognitively (despite having a maths & physics PhD background.. makes me very sad), although have tried to follow parts of this. Is it the case that @Chris Ponting is a member of the UK biobank? And his data does not support the paper that Dr Phair published? Why is this? And why do the OMF think this is @RDP ? and why does Chris ponting think this is?
I had been very excited and heartened by the paper that was published.

 

@RDP @Chris Ponting . I get lost following the explanation but this is what I found. Let me know if this makes sense?

Here is the coding MVV file that I think Chris is describing. I have extracted the variant described from the file
https://storage.googleapis.com/gnomad-public/release/2.1/mnv/gnomad_mnv_coding.tsv

Code:

39872934   8   ENSG00000188676   IDO2   ENST00000502986   Rescued nonsense   8-39872934-A-T   8-39872935-T-A   8-39872934-AT-TA   missense_variant   stop_gained   missense_variant   tAt/tTt   taT/taA   tAT/tTA   Y/F   Y/*   Y/L   None   HC   None   2.0   2.0   7.138981695650932e-06   0   2   280326.0   62658   280152.0   1.0   1.0   4.019163370952702e-06   0   1   248936.0   []   56546   248808.0   []   1   1   3.1904032669729455e-05   0   1   31390.0   []   6112   31344.0   []

The Multi-Nucleotide Variant in the file is described on this web page
https://gnomad.broadinstitute.org/variant/8-39872934-AT-TA

If you go back to the original variant page you will see that this MNV seems to apply to only two individuals in the database
https://gnomad.broadinstitute.org/variant/8-39872935-T-A

So if it is only two individuals that have this MNV I think we can say that for most people the STOP is valid. Am I right?

 

RE: "So if it is only two individuals that have this MNV I think we can say that for most people the STOP is valid. Am I right?"

Yes, you're correct: only two individuals, I hadn't spotted this. Thank you.
So for nearly all people that have the variant it does, indeed, truncate the protein. Thanks for putting me right on this.

Nevertheless, people who carry this variant are not more likely to be ME/CFS cases than controls, according to UK Biobank data.

 

If anyone is interested you can search by variant with the "Phewas" search option and you will get a table you can sort by p-value for that variant. I couldn't see any phenotype that was relevant for all four of the 4 "damaging" IDO2 variants the database supports. Might be of interest to @RDP

e.g. here is the phenotype table for IDO2 Y359STOP variant rs4503083. You will have to sort by p-value yourself by clicking on the down arrow next to "p-value".
http://geneatlas.roslin.ed.ac.uk/phewas/?variant=rs4503083&representation=table

 

I'm obviously not either of the two tagged people in your post but thought I'd try and help out.

No, to the best of my knowledge he isn't a member of the UK Biobank (not to be confused with the UK ME/CFS Biobank). He does, however, have access to the data from the UK Biobank and has experience in this area.

For background, from https://www.ed.ac.uk/profile/chris-ponting, "Professor Chris Ponting is Chair of Medical Bioinformatics and a Principal Investigator at the MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine."

Correct. The data from the UK Biobank shows that, for those people who self-reported a diagnosis of CFS, they were no more likely to have the IDO mutation than anyone else in the Biobank cohort.

Well, first we need to bear in mind that the paper is a proposed theory that needs to be validated. But I believe I'm right in saying that the authors report that, of 20 people tested by the OMF, all 20 are positive for the IDO mutation.

I won't try to answer the question why there is this seeming difference between the Biobank data and the metabolic trap theory, I'll leave that to the experts to figure out.

Hope this helps.

 

As far as I am aware OMF have at this point tested 65 patients and all test positive for IDO mutation.

 

For the fun of it I took a look at the high p value phenotypes for IDO1 and IDO2. Not sure it means much for the metabolic trap theory, but it was fun to play.

@RDP This IDO2 variant looks interesting in the UK Biobank data. It is not a variant you mentioned before, maybe because of low MAF of 0.01%
https://gnomad.broadinstitute.org/variant/8-39847306-C-T

This is the data in the UK Biobank for phenotype "B98 Other specified infectious agents as the cause of diseases classified to other chapters" for variant rs199869245. There were 918 cases in the database.

Code:

Variant              Chr  Position    beta              pv                MAF
rs199869245     8     39847306   -0.025354   6.3966e-09  0.0001173

Source : http://geneatlas.roslin.ed.ac.uk/search/?traits=82&variants=rs199869245

Given the low MAF I don't know how reliable this is but it does meet the -log10(pv)>8 criteria and infectious agent is interesting


Also I took a look at IDO1 variants in case that is of interest to you. This is the data for phenotype "O00-O08 Pregnancy with abortive outcome". There were 3984 cases in the database.
https://gnomad.broadinstitute.org/variant/8-39778730-C-T

Code:

Variant         Chr    Position       beta            pv                MAF
rs11995570   8       39778730    0.027008    9.1014e-09  0.00152519
rs10113525   8       39775845    0.025189    9.5611e-08  0.00148379
rs7840765     8       39776922    0.022857    6.6065e-08  0.00184858

http://geneatlas.roslin.ed.ac.uk/search/?traits=254&variants=rs10113525+rs11995570+rs7840765

 

Or maybe a series of adverse contributing factors, only one or some of which might be metabolic traps. Possibly a metabolic trap mechanism could be the final straw pushing people into the severe state. This makes me ponder the possibility that severity might not simply be a smooth continuum of severities, but maybe discontinuous stages, depending on whether a given mechanism kicks in or not.

 

I must say, I found this rather refreshing to read in the paper. Having become so tired of people trying to convince us that completely implausible and unhelpful ideas about ME/CFS are true, the suggestion that the theory has a low probability of being right actually increased my confidence in the authors and therefore, paradoxically, my perception of the probability of the theory being right.

A few more questions:

– Am I correct in thinking that a metabolic trap could also exist with no predisposing damaging mutations in enzymes or transporters? After all, even if this theory is correct, most people with a broken IDO2 will never develop ME/CFS, so presumably the same trap could theoretically have existed if everybody had a damaging IDO2 mutation – or even if IDO2 had never evolved.

– I understand that the authors looked for common mutations in human enzymes because of the existence of epidemics but, if it is assumed that there are predisposing genetic factors in ME/CFS, is it not at least as likely that the predisposing mutations would be in the genes which affect the probability of cytosolic tryptophan concentrations rising above the critical threshold in the first place rather than genes which determine whether or not someone falls into the trap when the critical level is reached? After all, if the theory is correct, it would to be the concentration of cytosolic tryptophan which is the most significant factor in determining whether or not someone develops ME/CFS.

– This makes me wonder, how could a metabolic trap be identified if there were no predisposing mutations in any of the enzymes or transporters?

– @RDP Can you tell us whether the testing of the theory based on metabolic tracer kinetics (as described in your paper) is happening, or likely to happen?

– If this type of metabolic trap is shown to be theoretically possible, would that mean that it is also probable that it exists in at least some people? Are there currently any good reasons to think that the trap does not occur in anyone?

– If it transpired that this (or any other) metabolic trap only applied to a small sub-group of people with ME/CFS, is there a danger that it could be wrongly rejected based on genetic analysis without getting to the stage of testing with tracer kinetics analysis?

– If IDO2 is broken in most people without it causing illness, is there any suggestion as to why the enzyme evolved, and why it appears to have subsequently been abandoned by evolution? Could it be that elevated cytosolic tryptophan concentration levels were once much more common for some reason?


Thanks again to Chris Ponting and Robert Phair for taking the time to respond to questions. Great to have you both involved with research and engaged with patients.

 

Check out Craig Robinson's maths catastrophe theory as applied to ME

 

I wonder if a trigger in susceptible individuals could be a drug or even supplement that people commonly use when they get an acute illness like the flu to alleviate symptoms.

I just googled, acetaminophen and tryptophan, just for fun, and this article came up. Disclaimer - I have no idea how factual it is, or if it's relevant to this discussion, but the authors say they have found that acetaminophen can cause falsely low serotonin levels in the test they perform, and that patients should therefore be advised to discontinue use before testing.

Interference of paracetamol (acetaminophen) with a commercially available high-performance liquid chromatography analysis of serotonin leading to falsely low serotonin levels

Albrecht Pfäfflin*, Karsten Müssig*, Erwin Schleicher

First Published January 22, 2009 Research Article

https://journals.sagepub.com/doi/full/10.1258/acb.2008.008116


If I'm remembering correctly that tryptophan is the precursor to serotonin, I wonder if acetaminophen could inhibit the breakdown of tryptophan to serotonin somehow?

And if someone had the IDO mutations and they were also a slow metabolizer of acetaminophen, could that make them more susceptible?

Please note, this is just a random thought, happy for someone to point out that it's nonsense. I was wondering if maybe even something simple like vitamin c could affect tryptophan levels in some way since so many people take it when they have the flu, I just haven't got around to googling that one yet.

 

Thank you Trish. I do wonder then (and I’m sure this has been talked about already, sorry I just can’t read/understand all the replies),1) could it be the case that some of the people reporting M.E/CFS in the UK biobank may be misdiagnosed or not reporting as accurately as could be done if specifically looked at by an ME physician. And in the UK, where I am, I do think of a few people who I believe have been misdiagnosed..I think it’s also in the way CFS & Medically Unexplained Illnesses (MUS)/Persistent Physical Symptoms (PPS) is diagnosed here and the criteria isn’t very strict. Although for that to affect the biobank data, I’m sure that number of misdiagnosis would have to be quite big. So what’s you said about severely ill patients and having extra metabolic traps could be something too.


2) I suspect that those diagnosed with severe M.E will be far more likely to actually have M.E (as in the OMF study). which is why I’m happy the OMF is using severely ill patients predominantly in many of these studies and hypotheses. Is there a way for the biobank to do the same? Severe meaning predominantly bedbound and very severe meaning unable to leave bed at all and requiring 24 hour care I think are the (loose) definitions.

Thank you very much Andy. Very sweet of you

 

No, I don't believe they can, but the proposed GWAS study that Chris would be part of would be looking, in part, at samples stored at the UK ME/CFS Biobank which, as they visit patients in their home, have many samples from severely affected patients. Should it actually go ahead, this then would provide data to support, or not, the metabolic trap theory.

 

Ah! What does GWAS stand for? That definitely sounds like a better way of providing data - visiting severely ill people in their homes.

Edit: brain fog so I got confused. Just editing the part below.

I used to think the UK biobank and the ME/CFS biobank were part of the same thing. Does the ME/CFS biobank not have data on severely ill patients and could they not use that data? (You’ve just answered that in the above reply so please ignore this part).

I’m still a little confused though as to why we have two biobanks!

 

Actually over 600.

 

Ah, yes, looks like it's less than the 600.

https://cureme.lshtm.ac.uk/researchers/accessing-the-biobank/

200 still seems a bit low but it's not the full 600 obviously. I can get it confirmed if we need to know exactly.