Machine Learning-assisted Research on ME/CFS

cc : @JaimeS ,@ScottTriGuy , @Perrier

Further to the hypothesis that ME/CFS originates from a Liver Injury caused by Viral Infection / Certain Medications / Toxic Substances here is more Research suggesting this hypothesis. The Research i present involves organophosphates and Depleted Uranium (DU). Please note that DU is thought to be the cause of Symptoms of Gulf War Ilness Syndrome researched by Nancy Klimas.

First, regarding DU :






We note : Bile Acid Metabolism disruption, involvement of CYP27A1 metabolites and LXR which can be seen in the Network Analysis graph (apart from FXR, LXR) created in 2017 and was also part of my presentation in EUROMENE :








Now, Organophosphates :






There was a Thread on Phoenix Rising regarding Organophosphates mentioning a plane crash in Holland. Among its cargo where organophosphates and depleted uranium was contained in the tail of the aircraft.


From Wikipedia :

and

Links :


https://forums.phoenixrising.me/ind...e-dichlorvos-and-mitochondria-toxicity.54754/

https://en.wikipedia.org/wiki/El_Al_Flight_1862#Cargo

 

cc : @ScottTriGuy @JaimeS


As a short update, i am providing a new potential target for further Research on ME/CFS. We have a very interesting finding regarding the NRG1 (Neuregulin 1) and its association with :

a) microglia activation
b) cholecystokinin
c) vagus nerve
d) AXL Receptor, member of TAM Receptors

Some excertps from research papers :

Microglia activation (relevant to Jarred Younger's findings)

Cholecystokinin (Patrick McGowan, Will De Vega ) with LXR (AI-assisted Research, unpublished)

AXL - part of TAM Receptors(AI-assisted Research, unpublished)

From my presentation at the EUROMENE, slide 50 where TAM Receptors are discussed :





I will contact Jarred Younger , Carmen Scheibenbogen , Will De Vega and Patirck McGowan in case this is of interest to them. In theory, it appears that dysregulation of TAM Receptors (due to Liver injury) may have deleterious effects in certain individuals.

Interestingly, NRG1 may be given intravenously so perhaps it may be considered as a potential treatment.

For references please visit the original post, here

 

It gets more interesting with NRG1 : Alan and Kathleen Light have identified NRG1 and Peroxisome proliferator A (PPARA - identified during AI-assisted medical research ) since 2016 :

On a paper named "Gene Expression Factor Analysis to Differentiate Pathways Linked to Fibromyalgia, Chronic Fatigue Syndrome, and Depression in a Diverse Patient Sample", a factor analysis table lists NRG1 and PPARA :

https://onlinelibrary.wiley.com/doi/pdf/10.1002/acr.22639


Thanks to @ScottTriGuy for pointing this out

 

Here is a new post where the connection of bile acids, thyroid metabolism, brown adipose tissue catabolism and thermogenesis are discussed. In a nutshell :


-Gene research targets : DIO2, UCP1, GPBAR1, PPARGC1A/PGC-1α
-Bile acid disruption / Liver disease results to increased rT3/TT3 ratio (this was found on a recent study to be a problem on ME/CFS patients)
-Thermogenesis due to cold exposure may be beneficial. This is confirmed possibly from the work of polish researchers, discussed in the presentation in this post from Dr Morten's presentation :

https://www.s4me.info/threads/dr-karl-morten-oxford-university-lecture-in-new-zealand-about-his-research-december-2018.7287/

A combination of DIO2, UCP1, GPBAR1 shows some symptoms that we've seen before (namely sensory discomfort, malaise, fatigue, pain) :







Liver disease may be responsible of altered rT3 /TT3 ratio :

EDIT : PPARGC1A was identified thorugh Network Analysis in 2017, see the following post where it appears as node "pgc1"

http://algogenomics.blogspot.com/2017/10/glucose-metabolism-fmo3-and-foxo1.html


Full post :


http://algogenomics.blogspot.com/2019/01/liver-bile-acids-and-thyroid.html

 

Thinking from the metabolic trap idea that other genes in the same function group may not compensate for mutations in others as has previously been assumed- what common genes linked to liver function might this apply to - high incidence, low penetration and substrate limiting? Would these instigate bistability?

Complete lack of GSTM1 is seemingly common- if there are other mutations/ compromises in its pathways and/ or other similar genes which are rate limited and don' t fully compensate what would be the effects?


Given liver performs so many key processes , and consumes so much energy to do so , I'm surprised that so little focus has been afforded it.

 

I am not sure what to say on this part unfortunately. However i do have a list of Genes (e. SP1, NR1H4, PTPN11 etc) that researchers could look at. Although i cannot give specific names, what i can say is that i was happy to find out that there was agreement in several candidate genes.

Exactly. It is the combination of genes that we want to be looking at, at specific pathways (according to the hypothesis)

Once again, i agree. We have a hypometabolic state and none of the teams have a Liver expert (=the main metabolic organ !!). What i am trying to say for quite some time now is that Liver disease can be missed in notoriously many ways.

Simple tests such as Total Bile acids have never been tested in any ME/CFS cohorts to this day.

I would also like to document the following. In the past 10 days i managed to put together a system that reads PUBMED articles for any given subject (e.g RBC Deformability) and pulls all genes that are found within these PUBMED posts. As discussed, i have a number of genes that -according to the hypothesis- are found to be important. I tried to match these genes along with genes found in several hundreds of thousands of pubmed articles related to inflammation, indoleamine/kynurenine and many other biological pathways /functions. At the time of writing it looks like this :



In other words most genes i found are related with liver disease, followed by inflammation, tyrosine kinase, cholesterol metabolism, cysteine catabolism etc.

EDIT : Only a certain subset of the genes i found are currently matched

EDIT 2 : Cytoskeleton is related to RBC Deformability

 

I just found out about Dr Broderick and his work. Way ahead from his time, used Network Analysis in 2011 for GWIS :

and

Link : https://www.elsevier.com/connect/how-text-mining-is-changing-the-way-we-tackle-chronic-disease


And an interesting Vimeo video on ME/CFS

 

Dear All,

I wanted to provide an update and some comments regarding this effort after having watched the NIH Confererence on ME/CFS.

a) It was very interesting to see that machine learning and AI is now being considered and used more than previously. I have seen excellent examples of use by Dr Phein and i couldn't agree more with him, especially about the use of text mining.

b) It was also very interesting to hear a comment from the audience addressed to Dr Montoya on using Network Analysis tools (Day 2 @7:38:52) to other research findings and try to see the whole picture. This is what we need.


I would like now to add some comments on why after watching the conference my confidence increased even more that Machine Learning may have had some answers already. More specifically :

====================================================================================================================================================

1) Dr Ron Davis discussed about the nanoneedle and for having a test that is able to differentiate with extremely high accuracy healthy controls vs ME patients. He mentioned about "exosomes" which are a type of extracellular vesicles. We see the role of GAS6, MER (=MERTK) and MFGE-8 :







Machine Learning has already confirmed this possible connection for GAS6, MERTK (note also FASLG = FAS Ligand) :



and MFGE8 :






More importantly, we also read about FAS Ligand (=FASLG) and TRAIL below in relation with extracellular vesicles :

and also note the "plasma membrane sensors" below :

TRAIL was mentioned by David Systom on Day 1, @5:08:00. FASLG has being mentioned by Michael Sikora at a previous presentation in 2018.

====================================================================================================================================================

2) Dr Mark Davis mentioned FGF21 as a potential target. Interestingly Liver and gallbladder are shown :




Source : https://videocast.nih.gov/summary.asp?live=31640&bhcp=1



====================================================================================================================================================

3) Mike Van Elzakker mentions TSPO Gene at day 2 @4:51:26




Relevant tweet on TSPO :






====================================================================================================================================================


4) Bhukesh Prusty

He mentions about DRP1 and its relation to mitochondrial fission. Interestingly, DRP1 is highly related to Endoplasmic reticulum stress (identified since 2015 as target of research) :

https://goldlabfoundation.org/presentations/drp1-a-link-between-er-stress-and-apoptosis/


There are is also an other potentially interesting relationship with SP1 transcription factor that i would like to bring to your attention without getting into details now.

====================================================================================================================================================

5) MAIT Cells : I posted some notes on the thread of the NIH Conference.


Actually there are more but i will not get into these. Given the above, I do strongly believe that we are getting closer. Things to watch for :

a) The test by Dr Ron Davis
b) the work by Bhukesh Prusty


I do not know what happened with the research at extracellular vesicles by Maureen Hanson. The AI system has identified the importance of EVs and for this reason i sent an email to Dr Hanson. This fact was also captured on a slide i presented at EUROMENE, where EVs are found on the first algorithmic run:

 

Thank you @mariovitali

For another interesting and informative post. I've printed it off to try and understand it better.

 

This may be a very important update. I would kindly ask you to PM me or report here regarding the following Pathogenic SNP :

rs7755898 , risk T, MAF=0.0025 (CYP21A2)

Although SNPs are not my area, this deserves to be looked at : In my cohort of 71 individuals :


-78.87% did not have this gene analysed
-19.72% individuals were found to have the risk allele.
-1.41% did not have the risk allele

Interestingly, CYP21A2 is part of the RCCX Module Theory :

https://me-pedia.org/wiki/RCCX_Genetic_Module_Theory

But this is where it gets interesting : CYP21A2 is related to Cortisol and Aldosterone (production of both is disturbed in ME/CFS patients) ,SRD5A2 (related to Post-finasteride syndrome) but also SP1, MERTK, Phagocytosis, LXR and many other topics that Machine Learning identified and were also discussed here.

Some excerpts from the post on my blog, linked below, noting SP1, EGFR, CYP21A2, MERTK :


"It is reported that many genes such as EGFR, have unequal copy numbers of the two alleles, particularly when one allele carries mutations, hyper- or hypomethylation.24 Many mutations occurring in highly methylated CpG dinucleotides in CYP21A2 lead to reduction of 21-hydroxylase activities, such as P30L, R356W and R339H.25–27 However it is unknown whether the copy numbers have changed for Q318X mutation because we are unable to detect the copy number variation. In addition, the allele containing the variants may not be normally bound by the transcriptional factors, such as Sp1 and SF-1"

and

"Cd36 deficiency led to reduced expression of phagocytosis receptor Mertk and nuclear receptor Nr4a1 in cardiac macrophages, the latter previously shown to be required for phagocyte survival. Nr4a1 was required for phagocytosis-induced Mertk expression, and Nr4a1 protein directly bound to Mertk gene regulatory elements"

Post : http://algogenomics.blogspot.com/2019/04/cyp21a2-target-with-many-potential.html



Tweets i made on EGFR, SP1, MERTK, LXR can be found in my twitter account.

EDIT : Based on the above, the implications of individuals having reduced functioning in both CYP21A2 and in any of SP1, CD36, NUR77/NR4A1, NR0B1/DAX1 and SF1 should be further investigated.

 

According to this SNPedia page 23andMe miscalled rs7755898
https://www.snpedia.com/index.php/Rs7755898

However I checked the "Lilly Mendel" reference files and could not find this SNP. @mariovitali is your data 23andme?

 

Thank you @wigglethemouse for help on a subject that i do not know. Yes, most of the data are 23andme.

EDIT : From DBSNP :


Gene:
CYP21A2 (GeneView)
Functional Consequence:
coding_sequence_variant,stop_gained
Clinical significance:
pathogenic
Validated:
by frequency,by cluster
Global MAF:
T=0.0001/23 (GnomAD_exomes)
T=0.0025/70 (GnomAD)

 

There is a bit more info here
https://www.snpedia.com/index.php/Congenital_adrenal_hyperplasia

 

Right, false alarm then with that SNP. Thank you !

 

There is some public 23andMe data posted here if it helps.
https://my.pgp-hms.org/public_genetic_data?data_type=23andMe
However there are at least 5 different versions of chips, and I couldn't find which one checks for that SNP.

 

Thanks !

Looking at one file who had this SNP it is Version 4. I will check more files to get their versions and post it here.

 

Here is a quick update, given the latest debelopment with CCI.
cc : @wigglethemouse , @ScottTriGuy


@Jeff_w , @JenB :


As you know i have been trying to identify what causes ME/CFS using Machine Learning methods. Given your recoveries please read below how it all may be connected along with the Liver :


1) Here are 4 algorithmic runs that i performed just today :















It is normal to have different results but here is the pattern that i see :

a) We observe that Liver Disease and brainstem are ranked in the top positions

b) We observe that acetylcholine is ranked high

c) We also see vagus nerve in two of the algorithmic runs

d) The following article, mentions Vagus nerve, Brainstem, CCI and ME/CFS

http://simmaronresearch.com/2019/03...chronic-fatigue-syndrome-the-vanelzakker-way/


Here is where the liver comes into play :

Link : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5788206/


e) Interestingly, we have among the algorithmic results entries for Cholecystokinin, ghrelin and leptin. All of these are related with the Vagus nerve :


https://www.frontiersin.org/articles/10.3389/fpsyt.2018.00044/full


f) We have seen cholecystokinin before in ME/CFS. From Will de Wega's paper cc : @ScottTriGuy :

g) We also know that ME/CFS patients have impaired bile acid metabolism which of course impairs fat metabolism.

Observe below how fats stimulate the vagus nerve. Acetylcholine is there, sepsis is there (recall that sepsis is on the results and Ron Davis mentioned a Sepsis-like state)



Source : https://www.researchgate.net/publication/7536881_Fat_meets_the_cholinergic_antiinflammatory_pathway




@Jeff_w

Could you show Dr. VanElzakker this post please? I strongly believe that Liver is an important part for a significant amount of patients along with the Vagus nerve.

 

@Jeff_w

Could you show Dr. VanElzakker this post please? I strongly believe that Liver is an important part for a significant amount of patients along with the Vagus nerve.[/QUOTE]

@mariovitali - super interesting that you've connected more dots of this puzzle.

Did you mean to ask @Jeff_w to show it to his doctor, Dr David Kaufman?

 

Sure, it may be Dr Kaufman as well. My assumption was that @Jeff_w might already be in contact with Dr VanElzakker also.