Jack | amatica health Profile picture
Aug 6 22 tweets 4 min read Read on X
🔬Simplified breakdown of the Decode ME results:

DecodeME identifies 8 gene regions linking immune response, mitochondrial energy control, and brain-cell signalling to ME/CFS

Genomic evidence the disease is biological.

Let’s breakdown everything in depth 🧵 Image
Study Cohort:

15,579 people with doctor-diagnosed ME/CFS + 259,909 UK Biobank controls (no ME/CFS).

85 % were women, average age ≈ 52 y.
How much is genetic?

Common SNPs explain = 9.5 % of overall ME/CFS risk (heritability on the liability scale).
   
For comparison:

- asthma = 10 %
- arthritis = 12 %
- type 2 diabetes = 13%
   
So ME/CFS is typical for complex diseases when you look only at common variants.
Key finding:
8 DNA regions change risk a little (odds-ratios ~1.08 ↑ or 0.93 ↓).

OR 1.08 = 8 % higher odds of developing ME

OR 0.93 = 7 % lower odds of developing ME

Multiple genes stack up to increase or lower risk.
Main genes & plain meanings:

- RABGAP1L (helps cells expel germs)

- BTN2A2 (activates a special T-cell)

- FBXL4 (keeps mitochondria healthy [energy])

- SUDS3 (controls brain immune cells)
- OLFM4 (tones down neutrophil bug-killing)

- CCPG1 (cleans stressed ER parts)

- CA10 (shapes nerve-to-nerve contacts)

- ARFGEF2 / CSE1L (manage TNF-α, an inflammation signal)
Where do these genes matter most?

A tool called MAGMA (it groups DNA signals by gene and checks which tissues use those genes) shows they’re used most in the brain. So the genetic clues link ME/CFS to the nervous system as well as the immune system.
Does infection matter?

Yes. In people whose illness began after an infection, the OLFM4 signal is much stronger; it’s absent in non-infection cases.
Male vs female DNA effects?

Variants act equally in men and women; male-only analysis lacked power but key female hits (CA10, ARFGEF2) still showed the same direction.
Immune insights:

HLA allele DQA1*05:01 was slightly protective (less common in patients). HLA genes help immune cells recognise threats.
Overlap with other diseases?

- The CA10 region is shared with multisite chronic pain (high probability it’s the same causal SNP).

- None of the eight regions share causal SNPs with depression or anxiety studies.
So will this help identify treatments?

When a disease-linked gene pinpoints a process (e.g., TNF-α release or mitochondrial upkeep) drug projects aimed at that process have higher success rate of projects without genetic support.
Example 1 - Inflammation angle

DecodeME noted a region with the genes ARFGEF2 / CSE1L that regulate how cells package and release TNF-α, a key inflammatory signal.

Existing anti-TNF drugs (used in rheumatoid arthritis & Crohn’s) could now be tested for ME/CFS.
Example 2 - Nerve-signalling angle

Another hit, CA10, shares the same causal variant with multisite chronic pain.

CA10 affects how nerve cells talk to each other. Compounds that fine-tune this synaptic pathway (already explored for pain) are now candidates to check in ME/CFS.
Will we be able to expand on insights with our new @amaticahealth RNA seq test?

If a person carries a “risk” or “protective” version of immune-expressed genes like RABGAP1L, BTN2A2, OLFM4, ARFGEF2, CSE1L we can see if their RNA level go up or down.
This allows us to see how the generic variant is impacting the functioning of the system.
Example:

If a risk DNA drops RABGAP1L RNA (weaker bug-clearing) or boosts OLFM4 RNA (stronger neutrophil brake), the up/down shift shows whether that variant turns immune defences down or inflammation up
More information on RNA-seq and the rest:

amaticahealth.com/me-cfs-long-co…
So overall very much what we expected from the study.

Risk factor genes that relate to immune system, mitochondria, and nervous system function.

The necessary next steps now are to determine if these alterations cause functional changes that can drive the disease.
Or are they simply just a ‘trigger’ risk.

I will come back to the Decode ME findings in a few months when we have our RNA data to see if we can confirm any changes in gene expression within these similar systems (TNF-a signalling etc)

And track against disease profile.
I will also do some more breakdowns on the exact genes and their prevalence in other diseases, if known, over the next few weeks.
Study pre print link:

pure.ed.ac.uk/ws/portalfiles…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Jack | amatica health

Jack | amatica health Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @JackHadfield14

Jul 30
🔬Simplified Break down of the recent AI paper that was circulating by @DeryaTR_ & co

Researchers used AI + multi-omics to decode the biology of this misunderstood disease.

Similar approach to what we will be doing @amaticahealth

So what did they find? 🧵 Image
The team followed 249 people over 4 years - 153 with ME/CFS, 96 controls.

They collected gut microbiome, blood metabolites, immune cell profiles, lab tests, and symptom reports at each visit.

Then trained a deep neural net (BioMapAI) to map it all.
Microbiome:

Huge drop in Faecalibacterium prausnitzii and other butyrate-producing bacteria.

Rise in bugs that ferment tryptophan and benzoate, which produce pro-inflammatory toxins.

Evident shift away from microbial balance.
Read 9 tweets
Jul 15
One of the most interesting trends in our recent @amaticahealth neuro-immune ME/CFS & Long COVID subgrouping data:

A Renin‑Angiotensin - neuro inflammation/injury axis signal.

And we aren’t the only people who have found this.

🧵👇🏻 Image
Image
Image
Cluster 3 individuals with high brain‑injury markers - S100B (astroglia) & NfL (axons) - also had higher plasma angiotensin II (Ang II).

Ang II is a vaso‑active peptide that *drives* endothelial + neuro‑inflammation via AT1‑ROS pathways.
Other studies:

- post‑COVID study showed AT1‑autoantibodies ↑ correlate with NfL ↑

frontiersin.org/journals/immun…
Read 8 tweets
Jul 11
Interesting paper - simple breakdown 🔬

At rest, the ME/CFS fluid had extra serine (a protein building block) and less of a folate-type vitamin called 5-MTHF.

That hints the brain’s ‘recycle & repair’ chemistry is off-balance.
You can see the hold up as:

- The middle-step product sarcosine is high

- Later products dimethyl-glycine and choline are low

Hints that the system stalls somewhere in the middle.
Other clues show the brain rerouting energy:

compounds like creatine and trans-aconitate are higher than normal, suggesting a detour rather than a full-on fuel shortage.
Read 7 tweets
Jul 10
Breaking down the meaning behind some of the markers used in these clusters.

The main drivers of the cluster 3 neuro inflammatory group are NEFL & S100b.

What do elevated NEFL & S100b imply?

🧵 Image
Image
Elevated NEFL (neurofilament light chain) levels = sign of neuron damage, especially to axons.

It leaks into blood/CSF when brain or nerve cells are injured. Image
High NEFL isn’t disease-specific.

It shows up in MS, ALS, Alzheimer’s, stroke, trauma, and more. It reflects how much and how fast damage is occurring.
Read 9 tweets
Jun 24
3 Biological Neuroimmune Subtypes in Post-COVID & ME/CFS 🔬❕

We mapped our @amaticahealth post-COVID + ME patients into three distinct biological clusters using Neuroimmune markers

Cluster 1; mitochondrial stress
Cluster 2; Non inflammatory
Cluster 3; Neuro inflammatory

🧵 👇🏻 Image
How we did it:

• Serum & Plasma biomarker panel → neuro, immune, RAS & neuro mito markers

• Unsupervised Euclidean clustering → C1, C2, C3

Markers used for clustering:

- NEFL
- S100b
- PINK1
- DRP1
- BH4
- Serotonin
- Rock1
- Rock2 Image
Image
Image
🔋Cluster 1 - Mitochondrial-Immune subtype

• PINK1 ↑↑ (mitochondrial recycling)
• ROCK1 ↓ (cytoskeleton/endothelial)
• ACE ↑, Ang-(1-7) ↓ → low protective RAS
• TWEAK & HIF-1α ↑ → inflam/hypoxia

Markers; No major neuro injury - some general inflam - mito stress high Image
Read 8 tweets
Jun 11
So there was a lot of talk recently about a paper discussing COVID-19 micro-clotting.

I thought I’d break it down in simple language.

The paper looked at why tiny blood vessels get blocked in severe COVID-19. They discovered, it is not the usual clots we hear about. 🧵👇🏻 Image
Inside capillaries, the “wall” is a layer of endothelial cells. When these cells die suddenly, they leave rough patches that blood cells can stick to.
The death process is called necroptosis - a kind of self-destruct that bursts the cell open.

Somewhat like a popped balloon leaving bits of rubber behind.
Read 12 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(