Machine Learning-assisted Research on ME/CFS

I wanted to provide you with a quick update. (cc : @Hip , @JaimeS )

We have two new targets as suggested by Machine Learning and Network Analysis methods : PTPN11 and EGFR.

Here is a network analysis graph that was generated using BioGrid :



We note the existence of GRB2, SOCS3 and EGFR. Intrerestingly PTPN11 was found to be related to Bile acids metabolism and Intestinal integrity.

@Hip you will recall our discussion about SOCS3 and its relation to Leptin :

https://forums.phoenixrising.me/ind...e-treatment-for-cfs.37244/page-82#post-736485


More in my blog post here : http://algogenomics.blogspot.com/2018/08/the-multiple-roles-of-ptpn11.html

 

SOCS1 and SOCS3 both put the brakes on in the immune system. Coxsackievirus induces SOCS1 and SOCS3 to evade immune response. Several other viruses also do this. See here. Pioglitazone lowers SOCS3.

 

@mariovitali, do you know if and which role the alkaline phosphatase (AP) could play here?

In which way would/would not your model be influenced if alkaline phosphatase is low/high?

Does someone know, is it correct that ATP is needed to activate/deactivate AP? On the other hand, AP is used to hydrolize ATP to ADP to AMP (see e.g. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2254479/#!po=8.92857, p.12)?

Would it follow that if ATP metabolism is impaired that the activity of AP is decreased, and would this be seen as low AP in the blood? The other way round, if AP is low, could this lead to an impairment in or a non-optimal ATP metabolism? And what would a high AP mean wrt. ATP? Normally, if there are liver problems, AP can be high. Does one have to distinguish tissue specific/non-specific AP here?

 

Sorry , I do not know since the way these models work is entirely different.

 

Has any CFS patient had a liver transplant? Anything in the literature or cases from CFS physicians?

 

I tried a quick search of "cfs liver transplant" at https://www.ncbi.nlm.nih.gov/pubmed/ and didn't find anything that looked significant.

 

Well, according to this webpage "Liver disease can be a complication of inflammatory bowel diseases (IBD), such as ulcerative colitis (UC) or Crohn’s disease."

Some of the microbiome work done so far has shown similarities in diversity between ME/CFS patients and those with IBD, though I'm not sure if that reflected any actual IBD in ME/CFS patients.

One thing I do know is that chronic ulcerative colitis is a risk factor for liver bile duct cancer (cholangiocarcinoma). A close relative of mine had ulcerative colitis most of her adult life and died of bile duct cancer at age 60.

Oddly, I was diagnosed with chronic ulcerative colitis about three years after the onset of ME, but, after a month on Azulfidine, the "CUC" was simply "gone" and was re-assessed as a passing gut infection from something I must have recently consumed while out of the country. The problem was that I'd had increasing symptoms of this "passing" infection for a couple of years and had never been out of the country. The whole thing was so strange that I really wonder about its possible connection to ME.

 

Moderator note
This post and posts continuing this discussion have been moved from this thread.

I wanted to have a closer look to the "Metabolic Traps" Hypothesis being presented by Dr Phair.

To perform my Research, i use million of PubMed articles discussing Medical Topics that are either directly or indirectly related to ME/CFS. When a new Research target or finding comes up (e.g cardiolipin antibodies were found) , the first thing to do is to query the topic among millions of articles and see the "hits" that this gets, much like Search engines do.

Next, i am trying to see whether this Topic has a "central" role to a Network of Topic Associations. An example of this approach was presented by Wenzhong Zhou at the Symposium.

Both "Indoleamine" and "Kynurenine" had several hits among the 709 Medical Topics i am looking at so these are -potentially- interesting targets, definitely something to look at especially since they are related with many interesting Topics such as Inflammation, T-Cells, NADH, Serotonin, Melatonin and many others.

 

One more round of algorithmic runs has been completed, i added several new targets so that they are scored for their relevance to ME/CFS.

Two of these targets were kynurenine and indoleamine given their role as suggested by Dr. Robert Phair (Tryptophan was already in). A first pass i made confirmed that these two targets had good chances to be relevant however i wanted to use all of the techniques applicable.

Below we see kynurenine being picked up by Machine Learning :







However we have a lack of relevance for these targets as suggested by Network Analysis and some other techniques. In other words, given the data used, the techniques and the methodology that i use, some relevance of the Metabolic Trap to ME/CFS is suggested ( but not a strong one).

It is also interesting to note the complete lack of Tryptophan among the results. Tyrosine on the other hand gives many strong "signals" in some of the methods used.

 

@mariovitali I hope you are in contact with Dr. Davis

 

@mariovitali - I don't have the knowledge to understand your posts re:machine learning I'm afraid, but it's clear you have put in lot of work and passion into it. Do you intend to try and publish any of this?

 

@Sunshine3

Dr Davis is aware of this work since February 2017. The techniques used were able to identify the involvement of Pyruvate dehydrogenase, dysregulation of Phospholipids and dysregulation of Bile Acids before any other Research work.

@Skycloud
My main effort is to convince Researchers to look for specific markers given this Theory. I have preliminary evidence of ME/CFS patients having Liver fibrosis which is quite significant (Dr Davis is also aware of this and many other Researchers in the field such as Maureen Hanson, Derya Unutmaz and Dr Robert Phair) . Despite extensive testing on ME/CFS patients no such test has been performed on a large scale. The same applies for a simple test called Total Bile Acids. Composition of Bile acids is also of high importance (Hypothesis).

A Swiss practitioner managed to get one of his patients to have a Fibroscan test. It was a Grade 2 Fibrosis (4 means cirrhosis). He will be performing more tests now (given that the first test he performed was positive) and i will know the results soon. All of this takes place with no funding of any kind.

After nearly 4 years, things may start be changing.


EDIT : Here is a very interesting paper for those of you who wish to follow up on the Metabolic trap :

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3779535/

 

Thank you for this hard work done pro bono. Thank you. How would one even treat liver fibrosis? I mean aside from transplant. Is it even possible. Very worrisome problem....

 

Shouldn't people who have ME for > 20 years have bad liver problems that are detectable via normal lab tests? Where's my error in thinking? Does this kind of fibrosis not progress/develop further?

 

This is a very logical question ***EDIT : but it makes the assumption that Liver Disease can be detected with blood tests. This is not the case and only Liver biopsy can rule out Liver disease***. I am not suggesting that all ME/CFS patients have Liver fibrosis or Liver Disease. But there are many alarming signs that some ME/CFS patients do have problems.

- Unfortunately i cannot submit here the numerous entries posted by several members on a specific FB Group, all of them suggesting Liver / Gallbladder issues. Many of them were shown in my presentation. The same applies from numerous conversations on Phoenix Rising. Why do many ME/CFS Patients cannot tolerate any kind of medication? Why they cannot tolerate alcohol? Why did Professor Hanson found several markers suggesting Hepatotoxicity and impaired Liver function?

-I've shown a particular case of a woman in her 40s with Liver fibrosis towards cirrhosis. She never found her Liver enzymes elevated but was unable to tolerate the slightest amount of medication. Upon looking at the Fibroscan, her hepatologist wouldn't believe the results and suggested she takes a second test at a different lab. The results were similar and her hepatologist would still not believe them (...) I discussed with one Researcher at EUROMENE regarding this and how was it possible that such Liver damage does not show in Liver tests. The Researcher told me that after significant Liver damage , it may not be even possible that certain markers can be raised.

-On a questionnaire i submitted having 80 responses, 22.1% of the respondents suggested that they have found elevated Liver enzymes at some point.

- One Researcher told me that there are molecules found in the blood of ME/CFS patients suggesting there is "fibrotic tissue" somewhere (this was not at EUROMENE).


We have to know if this is the case and we have to rule out this possibility.

 

But shouldn't at least some of the patients end up with liver failure?

 

@adambeyoncelowe Good question. Perhaps the process is very slow for the majority or for some reason stops after some time. I do not have any data regarding the progression of Fibrosis to patients that gave me their fibroscans.

In my presentation I mentioned a number of questions and some hypotheses to look at. Many of them are quite easy to check. Others require elaborate work because important details can be missed notoriously easily.

As an example , Maureen Hanson found metabolites suggesting impaired Liver function and /or hepatotoxicity.i assume that using a BioBank these metabolites can be easily checked to larger cohorts.

This may be a missing piece of the puzzle at least for some patients.

 

I wanted to provide an update regarding this effort.

cc : @Andy , @Inara , @adambeyoncelowe @ukxmrv @Perrier @andypants

Just before and right after my presentation in London, i got in touch with Dr. Charles Shepherd of the ME Association. Unfortunately i did not have the opportunity to meet him during my visit in the UK but Dr Shepherd has always been replying to my e-mails and so the communication was there.

I sent to Dr. Shepherd information about the Liver Theory (Research Notes, Fibroscans, Patient comments pointing to Liver function etc) and he said that he would forward this Hypothesis to the MEA Trustees. I also sent him the presentation file i used at EUROMENE and given the latest presentations of the Symposium at Stanford, how i believe this Theory fits in with the findings of Jonas Berquist, Alain Moreau, Michael Sakora and Jarred Younger providing the specific slide number relevant to each of these presentations .


I followed up also after my presentation with Dr. Shepherd so that i know when the MEA Trustees would reach a decision or to have some form of feedback on how this Theory appears to them.


Dr Shepherd explained to me that this Theory may be eligible for funding but i will have to get a UK Liver specialist along. I replied that i will try to do that but in the meantime i could provide -free of charge- the software system to -hopefully- advance their Research efforts, helping put more pieces of the puzzle together at a quicker pace. I am expecting his answer to this but if someone knows a UK Liver expert i could talk to please let me know.


Things would turn out to be even more interesting. I watched a BBC Documentary about an ME/CFS patient named Emma Donahoe.


Let's see in the point where she talks about her regimen :

https://www.youtube.com/watch?v=XLPCuEdqIWY

So she takes spirulina, Fish Oils, D3 + K2, Vitamin B complex, Krill Oil and activated Charcoal. As yet another coincidence (?) Emma's Regimen and mine (which i showed at EUROMENE in Slide 63) have a lot of common supplements.

I will not get into the details on how Omega 3 Oils fit into the picture below (its through induction of certain Peroxisome Proliferators) but -interestingly- Krill Oil is rich in both Omega 3 and a Phospholipid called PhosphatidylCholine.

Looking into the Network Analysis chart, please find "Choline_Deficiency" in the lower part (centre) of the Network snapshot.







For those of you who have been following my posts on Phoenix Rising (e.g @JaimeS) , you will recall the several posts i made about Choline (which incidentally is important for proper Liver function). I had to stop Choline because i realised that it affected negatively my appetite so it is not part of my regimen anymore (and so it did not appear on the slide i presented at EUROMENE regarding my personalsied regimen)

 

I ended up not being able to have a single piece of food, it was not a good feeling at all. Upon looking to what may have happened I linked that incident with having too much acetylcholine.

I was getting warning signs for a couple of months before where I would see my appetite very slowly diminishing but I foolishly kept taking it.

I was taking Choline for a good 9 months non stop