Machine Learning-assisted Research on CFS

[mariovitali]

Just wanted to share a snapshot from a conversation on a CFS group.

The case is a 18-year old girl from the UK. Her CFS started with Urticaria all over her body.

So i tried talking to her in order to understand how she got CFS. She replied that she had taken no medication and thorough tests (including one for EBV) have been performed.

After reading the post about the Research on CFS, Post-Finasteride, Post-Accutane etc she comes back saying that she was taking Accutane for a year !

Here is the snapshot :

Of course i do not state that Accutane has caused CFS in this case. This is a hypothesis that has to be properly evaluated by Medical Researchers.


@JaimeS

How many cases of CFS sufferers becoming symptom-free -using a specific regimen- would you say that are enough to warrant further investigation to the hypothesis presented here?

 

[A.B.]

Of course i do not state that Accutane has caused CFS in this case. This is a hypothesis that has to be properly evaluated by Medical Researchers.

I was also taking accutane at the time of illness onset. I'm still ill 18 years later.

 

[JaimeS]

How many cases of CFS sufferers becoming symptom-free -using a specific regimen- would you say that are enough to warrant further investigation to the hypothesis presented here?

Very hard to say, @mariovitali .

Part of the issue is that when they get well, they understandably 'disappear' from public spaces like PR.

I have recovered an enormous level of function via self-hacking, but it would be impossible to call myself 'well'. Sometimes, I still crash. So I'm still here, hanging around, getting ideas. (Personally I'd still be hanging around because I am working on SCIENCE!!!, but the same cannot be said for everyone.)

If you're looking for some dubious sounding data, there's that oft-cited "5% recover" stat. However, early-diagnosis patients (4 mo of illness) may recover just fine. I'd be interested to see recovery stats on patients sick for over a year. New patients may have more of a chance of recovery than long-term patients.

I had a friend with post-viral fatigue -- we're talking months and months of what looked a great deal like ME/CFS -- but she recovered. Perhaps during those months her metabolomics 'looked' like an early-stage, acute patient. Then, she recovered.

But I know zero cases of long-term illness coupled with long-term recovery.

I know cases that look like recovery. People who return to work and get on with their lives.

Two years later they're back, maybe sicker than the last time.

Long story short (known as TL;DR in the world of the interwebs):

• Perhaps there are more recovered patients than we know of, but they leave 'patient-designated' spaces;
• Perhaps 'recovery' only appears to be recovery (patients have long-term remission before relapse) compounding the issue;
• There is a 5% recovery stat, but this may or may not apply to cases longer than 6 mo.

-J

 

[mariovitali]

@JaimeS

I cannot possibly disagree with your concerns regarding recoveries, especially since the same recurrence of symptoms has happened to me a couple of times when i got a mixture of CFS and PFS, back in 1998.

I had severe symptoms for about 6 months (if memory serves me well) and then i was fine for about a year and a half (again if my memory serves me well). Then Symptoms came back and over the period of many years i was getting worse year after year.

According to my Theory (which i still did not have the chance to disclose to any Researcher despite my efforts) it actually makes sense to have recurrences since the main culprit is not being taken care of. The question then is what is the main culprit and whether this has been taken care of with the regimen being given.

Considering recurrences of Symptoms, things are not so clear-cut and allow me to explain what i mean. Consider the following examples :

(1) Assume that you have a small group (n=5) of Individuals having problems for at least 3 years, having tried everything to feel better (from "Liver flushes" to Anti-depressants to Meditation) became fully functional within 3-4 weeks being on a specific regimen.

(2) An individual (female) having CFS/Fibromylagia for 24 years and after receiving a regimen for 2 months, she is able to function again, becoming 80% symptom-free (and getting slowly even better).


a) Would you agree that any regimen that is able to achieve the results discussed in (1) and (2) deserves at least to be considered for further investigation?

b) If these individuals do exist (i am one of them), who is the person that they should get in contact with for further evaluation?

c) I have no objection that this regimen is tried to individuals being selected by Researchers and only by them. No drug of any kind is used and everything used can be found OTC

Recall that at the end of my post at AlgoGenomics, hypothesis No 2 states the following :

Specific interventions according to an individual's DNA may significantly ameliorate or even reverse the Symptoms that are associated with the Syndromes discussed in this post

So i am talking about Symptom amelioration/reversal and not about Recovery, in full knowledge that for CFS/PFS etc it is not easy to talk about "recoveries". (This discussion has also reminded me that i need to correct my signature btw)

I have disclosed a number of Genes in my post (more coming soon). I would kindly ask from any Researcher you happen to know to consider them as hypothesis to their Research.

-As a side note : Several patients of CFS (but also PFS/Post-Accutane etc) complain that they do not get infections any more :

http://forums.phoenixrising.me/inde...cluding-q-and-a-with-dr-naviaux.46520/page-32

I believe that if a specific supplement is given to these patients they will most likely start getting viral infections.

I can e-mail to any interested Researcher this supplement and also explain why i believe this supplement achieves such result (if it indeed does so that is).

@All
Not disclosing this supplement to this forum is not about being secretive, it is about trying to be responsible and having this information available for evaluation by qualified Medical practitioners only.

 

[Jesse2233]

I had a friend with post-viral fatigue -- we're talking months and months of what looked a great deal like ME/CFS -- but she recovered. Perhaps during those months her metabolomics 'looked' like an early-stage, acute patient. Then, she recovered.

Jaime how long was your post viral friend sick and how badly impaired was she?

 

[JaimeS]

She was considerably impaired; it was almost exactly 6-mo. Not... "severe ME" impaired, but on the level of "minor" ME which is still pretty dramatic, symptom-wise.

-J

 

[Snow Leopard]

The problem with these sorts of attempts is the "garbage in = garbage out" phenomena. A lack of high quality underlying research makes it difficult to get good results. A lack of specificity of patient groups doesn't help either.

I have initially dismissed the involvement of MERtk since I expected that disruptions (eg autoantibodies targeting this receptor) would lead to clotting disorders, or certain cancers.

But for what it's worth:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4162177/

TYRO3, AXL, and MER are also implicated with the entry of certain viruses into the cell, so there is a possibility for post-infection syndromes. MER is also activated by EBV. http://jvi.asm.org/content/78/21/11778

 

[JaimeS]

a) Would you agree that any regimen that is able to achieve the results discussed in (1) and (2) deserves at least to be considered for further investigation?

b) If these individuals do exist (i am one of them), who is the person that they should get in contact with for further evaluation?

c) I have no objection that this regimen is tried to individuals being selected by Researchers and only by them. No drug of any kind is used and everything used can be found OTC

a) Sure! I would be wary, since many supplements and approaches seem to 'wear off', but I would definitely think that was interesting and worthy of further study.

b) There is no such single individual. I work at Stanford now, and you can talk to me. I can bring it to the group if I think it's meaningful. But you could probably talk to anybody on a research team and ask them. Or you could apply for a grant as an individual researcher, yourself. Run a (probably non-blinded) trial.

c) Everything being OTC is good news for patients who cannot afford expensive treatments.

Good luck!

 

[alex3619]

If you mention dozens of things (as done in that post) or hundreds (as done in that thread), odds are pretty good that you'll randomly hit on something that will be mentioned again in ME research.

[Satire] I would like to use Websters dictionary as a source of the theory to cure ME. Its got all the right words, we just have to put them together. As proof I would like to point out that just about all the papers published lately mostly use words found in the dictionary. [End Satire]

Heuristic tools like this can throw up possibilities, but they still have to be checked manually (at least at this stage) for plausibility and details.

The numbers game is also how quackery works. My hypothetical example has always been that you treat a thousand, a hundred improve, and ten are so convinced they give testimonials. A few go on to sell the therapy. When those ten to a hundred get together they all agree the treatment works. There, treatment proved! ... Not.

 

[mariovitali]

@Snow Leopard

Actually it is a combination of these Genes and others (Hypothesis). I agree with you on GIGO and there is also Confirmation Bias (among other things) that we should be aware of.

https://en.wikipedia.org/wiki/Confirmation_bias

The whole point here is to have a tool that makes informed hypotheses that may then be evaluated by Medical Professionals. If the hypotheses are indeed informed, well for this we should wait and see.


@JaimeS

Thank you and i agree with everything you say about your concerns. I do not wish to apply for any Grant for this purpose.

I would rather disclose all the hypotheses of my Research to a team rather one individual (as instructed to do so by my Lawyer), we can discuss the details if you wish. I think the process is quite simple : I will disclose a number of Genes and Biological Pathways that the tool identifies as important to CFS. I can also discuss a Hypothesis on why the human body gets stuck in this state of CFS/PFS etc. To be honest i would not feel very confident about doing this since i have no knowledge on Medical issues. But we can give it a try, even if it's one in a million.

The blending of "Man and Machine" is important here. I know nothing about Medicine but it would be really interesting to see what will happen if we have a tool that creates informed hypotheses on one hand and experienced Medical Researchers evaluating these hypotheses and guiding the tool on the other. Recall that the post in Algogenomics clearly states that one of the first things that needs to be done, is to identify why an Algorithm suggests some Topics. Are these Topics related to each other ? Which of these associations are known and which of them are unexpected? (and thus should be further researched).

@All

Machine Learning in Medical Research is not something new :

http://www.mlpm.eu/

http://clinicalml.org/research.html

 

[Snow Leopard]

The whole point here is to have a tool that makes informed hypotheses that may then be evaluated by Medical Professionals. If the hypotheses are indeed informed, well for this we should wait and see.

Sure, which is why I look at studies like this to see what effect inhibition of the receptor may have on humans:

https://www.ncbi.nlm.nih.gov/pubmed/20434586
See also:
http://www.nature.com/nbt/journal/v26/n1/extref/nbt1358-S1.pdf

The reason why I'm particularly focused on specific receptor function, rather than downstream enzymes is that for the same dysfunction to occur in cells all over the body simultaneously requires dysregulated signals outside of the cells.

 

[JaimeS]

@mariovitali -- I understand your concerns regarding someone selling or profiting off of your idea on which you have worked so hard. However, it's hard to get a meeting with a group of scientists without any explanation of what your idea is or why it should work.

 

[JaimeS]

Of course if you did forward it completely to OMF -- without the omissions we see here -- then it will eventually come to us @ Stanford.

 

[mariovitali]

@JaimeS

Sorry for my late reply!

I do understand your concerns that without a sufficient explanation things cannot move forward. I also did not forward the undisclosed version of my research to OMF.

One month ago i contacted the head of a Research team that researches the Post-Finasteride syndrome. The person who was responsible said that my Research looked interesting and asked for the specific SNPs that i believed should be looked at their cohort of PFS sufferers.

I replied that i could disclose everything as soon as we had a Non-disclosure Agreement (NDA) signed. The professor replied that they would verify if the mutations on genes i identify as potentially involved in PFS are so in their cohort of PFS subjects and that they would sign an agreement that *IF* a paper will be published on their findings based on my early evidence i will be a co-author.

Question: If you were in my position, would you move forward with this requirement as set forth by the responsible person of the Research team under these circumstances? My Lawyer said that i should not move forward and this is what i did.

As previously discussed, i do want to help, i want to do so asap and i am not asking for any kind of compensation. I will do anything i can to help out.

 

[Jonathan Edwards]

I cannot really follow what this is attempting to do but it seems likely that it is just picking up on what is currently trendy in biomedical science. What is trendy is nearly always wrong because once something is sorted it is no longer trendy because it is sorted. Trendy topics tend to be blind alleys. And researchers spend much of their time copycatting each other in order to be as trendy as they can.

 

[Valentijn]

Question: If you were in my position, would you move forward with this requirement as set forth by the responsible person of the Research team under these circumstances?

I really don't think you should bother, since you don't distinguish between SNPs which are capable of having an impact, and normal variants which are extremely common.

 

[Tunguska]

@mariovitali I appreciate what you're trying to do and I've been interested in what you did, but I thought the exact same thing as Jonathan Edwards and I place a lot more value on understanding the *biochemical (3 hours sleep) pathways nowadays. I think your method is somewhat unique but you're not the first to use meta-analysis in health issues and I noticed some of the people doing that I'd been following for years, ran into dead ends.

I really benefitted from your other thread that advised to try TUDCA. It's a powerful substance no matter how you reason its use. If it was cheaper and the supplements weren't all loaded with crap, it might be the best thing available.

I could be wrong but I was under the impression Vitamin K was of limited use for liver health?

Choline is extremely important for liver health in disease states. But this is not new information, Chris Masterjohn went deep into the old research a few years ago. The problem is that oral choline suffers from awful absorption/metabolism problems, and it leads to higher acetylcholine. Those two things can easily offset the liver benefits or prevent you from taking it. If you could just inject choline directly into the liver you could probably cure liver problems. The other vitamins are probably mostly there to support the choline processing [well and also blood sugar regulation].

 

[mariovitali]

@Tunguska
The thread on Unfolded Protein response is outdated, i have posted several times that findings (as hypotheses) have progressed since then and that the only thing remaining right now is to have Medical Researchers evaluate the hypotheses being generated by the Machine Learning Algorithms

@Valentijn

As previously discussed, i like the fact that all of our dialogs have been documented in this forum.

Before i begin, you also posted on this Thread by @Hip regarding Drugs linked to Mitochondrial Toxicity

http://forums.phoenixrising.me/inde...ne-prozac-xanax-tetracycline-metformin.51574/


You will recall that among the topics that have been discussed in the Unfolded Protein Response thread one of them has been ER (Endoplasmic Reticulum Stress). Now, please look the following parts of text from a paper (@JaimeS could be interesting to you and the team at Stanford as well ):

Recent studies suggested that endoplasmic reticulum (ER) stress could be an important event involved in drug liability, in addition to other key mechanisms such as mitochondrial dysfunction and oxidative stress.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4777263/

Then we have the following Figure :


https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4777263/figure/prp2211-fig-0002/


@Valentijn you will see topics -which you must have read in the Unfolded Protein response thread- such as

-ER Stress
-Accumulation of Misfolded Proteins,
-Glycosylation

and also on the list of abbreviations you will find PPAR. PPARa has been selected by the algorithms on the Node Diagram i posted on this Thread which is seen in a red rectangle here :

EDIT : PPARγ (mentioned in the paper i linked) is also shown on the Network Graph, forgot to annotate it, sorry

Looking forward to your comments.

 

[Valentijn]

Looking forward to your comments.

Nothing you've said addresses your insistence on treating all SNPs on a gene as if they're relevant to how it functions.

 

[mariovitali]

Nothing you've said addresses your insistence on treating all SNPs on a gene as if they're relevant to how it functions.

Great, Thanks !