Daniel sent us this one — he's asking about methylation, which gets thrown around a lot in health circles as a root cause of disease, and he wants to know what it actually means in plain terms. But the real question he's digging into is the connection to inborn errors of metabolism, or IEMs. Are those genetic diseases? Where does methylation actually fit into that story? Because these two concepts get tangled up constantly, and most of what you hear online collapses the distinction entirely.
This is one of those topics where the wellness industry has done a spectacular job of muddying the water. I want to flag something right at the top — by the way, today's episode is powered by DeepSeek V four Pro.
Alright, let's untangle this. Where do you want to start?
Let me start with methylation itself, because the term gets used as if it's one thing, and it's not. Methylation is simply the addition of a methyl group — that's one carbon atom bonded to three hydrogen atoms, C H three — onto a molecule. That's it. It's a tiny chemical tag. But where it lands matters enormously. When it attaches to DNA, specifically at what are called CpG sites where a cytosine sits next to a guanine, it typically silences gene expression. It's like putting a physical block on the gene that says "don't read this right now.
It's a dimmer switch, not an on-off toggle.
Or more precisely, it's one of several dimmer switches. DNA methylation is a core epigenetic mechanism — epigenetic meaning it changes gene activity without altering the DNA sequence itself. The underlying code stays the same, but whether and how much a gene gets expressed changes. And this is normal. Methylation is not inherently good or bad, which is the first thing most popular coverage gets wrong.
That's worth pausing on, because you hear "support your methylation" like it's a nutrient deficiency. As if more methylation is always better.
Right, and that's completely backwards. Abnormal methylation can go in either direction. Hypermethylation — too many methyl groups at specific genes — can silence tumor suppressor genes. That's a hallmark of many cancers. Hypomethylation — too few methyl groups — can activate oncogenes or cause genomic instability. The pattern matters, not some global "methylation level." You can't just take a supplement and say you're improving your methylation. You might be making things worse depending on the context.
When someone says methylation is a causative factor in disease, what they're usually doing is confusing a normal regulatory process with its dysregulation.
And the dysregulation is often secondary to something else. Which brings us to inborn errors of metabolism. Let me lay out what IEMs actually are, because the name itself is a bit of a historical artifact.
It sounds vaguely Victorian.
The term was coined by Archibald Garrod in the early nineteen hundreds. He was studying a condition called alkaptonuria, where patients' urine turns black when exposed to air. He realized it was caused by a missing enzyme in a metabolic pathway, and that it ran in families in a pattern consistent with what we now know as autosomal recessive inheritance. He called these "inborn errors of metabolism" — the idea being that the body is born with a chemical mistake.
These are definitively genetic diseases.
Every IEM is caused by a mutation in a gene that encodes an enzyme or a co-factor needed for a specific metabolic pathway. Collectively they affect about one in two thousand five hundred births. The mutation creates a block. Substrate A is supposed to be converted to product B by enzyme X, but enzyme X is defective. So substrate A builds up to toxic levels, and product B becomes deficient. That's the fundamental model.
The classification system — how are these organized?
By the pathway that's broken. You've got urea cycle disorders, organic acidemias, fatty acid oxidation defects, amino acidopathies like phenylketonuria, carbohydrate disorders, mitochondrial disorders, lysosomal storage disorders. Methylation is not a standard classification axis. But — and this is where the two stories genuinely intersect — there is a specific subclass of IEMs that directly involve the methylation cycle.
Okay, this is the remethylation piece.
Let me walk through the biochemistry quickly, because it's elegant. Your body has a universal methyl donor called S-adenosylmethionine, or SAM. SAM is the molecule that actually donates methyl groups to DNA, to proteins, to neurotransmitters — it's the methyl currency of the cell. After SAM donates its methyl group, it becomes S-adenosyl-homocysteine, which then gets converted to homocysteine. Homocysteine can go one of two directions. It can enter what's called the transsulfuration pathway and eventually become cysteine, or it can be remethylated back to methionine, which then gets rebuilt into SAM. That remethylation step — converting homocysteine back to methionine — that's the remethylation cycle. And it requires specific enzymes and co-factors, including vitamin B twelve and folate.
A defect in that remethylation machinery means you can't recycle homocysteine, SAM levels drop, and suddenly your entire methylation economy collapses.
And these are devastating disorders. Severe MTHFR deficiency — and I need to distinguish this from the common MTHFR polymorphisms immediately.
Let's do that, because I know that's going to be the thing listeners have actually heard of.
The common MTHFR variants, C six seven seven T and A one two nine eight C, are found in roughly twenty-five percent of the population. They're mild genetic polymorphisms. They might reduce enzyme activity modestly — think ten to thirty percent — but they are not disease-causing. They are not IEMs. The wellness industry has built an entire supplement empire around these variants, claiming they cause everything from fatigue to depression to infertility. The evidence for most of those claims is thin to nonexistent. And yet people are paying for expensive methylated B vitamins based on MTHFR testing that most geneticists consider clinically useless.
Because the variant is common and the enzyme still works.
Now, true severe MTHFR deficiency is a completely different entity. It's a rare autosomal recessive disorder where enzyme activity drops below twenty percent. These patients develop homocystinuria — extremely high homocysteine levels in the blood and urine. They present in infancy with acute neurologic deterioration, feeding difficulties, developmental delay, seizures, and a high risk of thromboembolism. Without treatment, outcomes are poor. This is a life-threatening IEM.
The gap between "I have an MTHFR variant and feel tired" and "I have homocystinuria and might have a stroke as an infant" is a canyon.
It is, and the conflation of these two things is arguably one of the worst science communication failures of the last decade. The RACGP, the Royal Australian College of General Practitioners, published a piece back in twenty sixteen saying MTHFR genetic testing should not be ordered routinely because it leads to unnecessary anxiety and supplement spending with no proven benefit. And yet here we are.
Beyond MTHFR, what are the other remethylation disorders?
There's methionine synthase deficiency, also called cblG, and methionine synthase reductase deficiency, cblE. These are defects in the enzymes that use vitamin B twelve to remethylate homocysteine. Then there are the combined remethylation disorders — cblC, cblD, cblF, cblJ — which are defects in cobalamin metabolism that affect both the remethylation pathway and a separate pathway involving methylmalonic acid. CblC is the most common of these, and it's still extremely rare. These kids are incredibly sick. They often have failure to thrive, microcephaly, seizures, visual impairment. The severity depends on how early they're diagnosed and how aggressively they're treated.
Newborn screening catches these now?
Many of them, yes. And that's been transformative. There was a study published in the Journal of Inherited Metabolic Disease in twenty twenty-two showing that pre-symptomatic diagnosis — catching these kids through newborn screening before they crash — is independently associated with significantly better neurodevelopmental outcomes. The difference between identifying MTHFR deficiency at day five of life versus day forty-five when the baby seizes is measured in I. points and motor function. It's night and day.
The methylation story in IEMs has two layers. Layer one is these remethylation disorders where the methylation cycle itself is broken because the enzyme is missing. That's direct causation. Layer two is something broader — what about IEMs that don't involve the methylation cycle? Does methylation still play a role there?
This is where it gets really interesting, and where a lot of the current research is focused. There was a review in twenty twenty by Rutten, Rots, and Oosterveer in the Journal of Inherited Metabolic Disease that lays this out beautifully. Their argument is that metabolic intermediates directly regulate epigenetic enzymes. SAM and S-adenosyl-homocysteine control DNA and histone methylation, acetyl-CoA drives histone acetylation, NAD plus affects sirtuin-dependent deacetylation. So there's this metabolic-epigenetic feedback loop. In any IEM, the primary genetic defect causes a metabolic disturbance — a buildup of something, a deficiency of something else — and those metabolic changes can secondarily alter DNA methylation and histone marks across the genome.
The methylation changes are downstream, not upstream. The gene mutation is the root cause. The metabolic block is the primary consequence. And the altered methylation patterns are a secondary effect that might modulate how severe the disease is.
That's the model. And it helps explain something that's been puzzling clinicians for decades, which is poor genotype-phenotype correlation. You can have two patients with the exact same mutation — say, in phenylalanine hydroxylase, the enzyme defective in PKU — and one has severe intellectual disability while the other has relatively mild symptoms. Same broken gene, different outcomes. Part of the answer may be epigenetic. The metabolic environment created by the enzyme deficiency alters methylation patterns differently in different individuals, and those epigenetic differences affect how other genes compensate.
That's a much more satisfying explanation than "we don't know," which is what the textbooks said for years.
It opens up therapeutic possibilities that didn't exist before. If the epigenetic changes are contributing to disease severity, maybe you can target those changes directly. The Rutten paper specifically discusses epigenetic editing as an emerging approach.
Walk me through what epigenetic editing actually means, because I've heard of CRISPR for gene editing, but this sounds different.
It is different, and the distinction matters. Traditional gene editing with CRISPR-Cas9 cuts the DNA and permanently alters the genome sequence. It's a one-way street — once you change the sequence, it's changed forever, and any off-target effects are permanent too. Epigenetic editing uses a modified version of CRISPR where the cutting enzyme, Cas9, is deactivated — it's called dCas9, dead Cas9 — and instead it's fused to epigenetic enzymes. You can attach a DNA methyltransferase to add methyl groups at a specific gene and silence it, or a demethylase to remove methyl groups and activate a silenced gene. The DNA sequence itself remains untouched.
You're manipulating the dimmer switch without rewiring the house.
Because epigenetic marks are naturally reversible, this is considered potentially safer than permanent gene editing. If something goes wrong, the epigenetic change can theoretically wear off over cell divisions, whereas a bad gene edit is permanent.
What would the application be for IEMs?
You could try to boost residual enzyme activity from a partially functional mutated gene by removing repressive methylation marks. Many IEM mutations don't completely abolish enzyme function — there's some residual activity, maybe five or ten percent. If you could epigenetically upregulate that remaining function to twenty or thirty percent, that might be enough to cross a clinical threshold and prevent symptoms. Alternatively, you could try to upregulate a compensatory enzyme in a parallel pathway. The idea is to work with what the patient still has rather than trying to insert a whole new gene.
That sounds elegant in theory. How far along is this actually?
Mostly cell culture and animal models at this point. Nobody is doing epigenetic editing for IEMs in humans yet. But the pace of development in the epigenetic editing field generally has been rapid. There are companies working on it for cancer, for neurodegenerative diseases. The IEM application is a logical extension.
The ethical dimension here is interesting too. If you can reversibly activate or silence genes without altering DNA, where's the line between therapy and enhancement? If you can boost enzyme activity in a PKU patient, could you boost enzyme activity in someone without PKU to make them metabolize phenylalanine even more efficiently? Is that enhancement?
How stable are these changes across cell divisions? If you epigenetically activate a gene in liver cells, does that persist when those cells divide? Do you need to re-dose? These are open questions. The Rutten paper raises them but doesn't resolve them, because the data isn't there yet.
Let me pull us back to something more immediately practical. You mentioned homocysteine testing earlier. Is that underutilized?
I think it is, especially in neurology. When an adult presents with unexplained neurologic symptoms — early-onset stroke, progressive cognitive decline, psychiatric symptoms, neuropathy — checking a plasma homocysteine level is cheap, widely available, and can point toward a remethylation disorder or a B twelve deficiency that's been missed. Many of these adult patients have been bouncing between specialists for years. The remethylation disorders are classically thought of as pediatric conditions, but milder forms can present in adolescence or adulthood. The diagnostic gap is real.
The treatment, if you catch it?
For the cobalamin disorders, high-dose hydroxycobalamin — that's injectable B twelve — plus betaine, which provides an alternative methylation pathway, and sometimes folate supplementation. For severe MTHFR deficiency, betaine is the mainstay because it provides a bypass for remethylation that doesn't require the MTHFR enzyme. These treatments aren't cures, but they can dramatically improve outcomes when started early. The tragedy is when the diagnosis is missed until irreversible neurologic damage has occurred.
Let me synthesize what we've covered, because this is a topic where it's easy to get lost in the biochemistry. Methylation is a normal regulatory process — adding a methyl group to DNA or other molecules to control gene expression. It's not inherently good or bad. Inborn errors of metabolism are genetic diseases caused by mutations in enzyme-coding genes that block metabolic pathways. Some IEMs directly disrupt the methylation cycle — the remethylation disorders. Those are rare, severe, and increasingly caught by newborn screening. But in the broader universe of IEMs, methylation plays a secondary role — the metabolic disturbance caused by the enzyme deficiency alters epigenetic marks, and those alterations may explain why patients with the same mutation can have different outcomes.
That's a clean summary. I'd add one thing: the MTHFR confusion is a cautionary tale. A common genetic variant with mild effects got marketed as a root cause of chronic disease, and millions of people are now taking supplements they probably don't need. Meanwhile, the actual severe MTHFR deficiency — the life-threatening IEM — is vanishingly rare and requires aggressive medical management, not a methylated B vitamin from a health food store. The gap between the popular narrative and the clinical reality is enormous.
It's almost a case study in how scientific nuance gets stripped away when a concept hits the wellness market. Methylation goes from "complex epigenetic regulatory mechanism" to "thing you need to optimize." IEMs go from "rare genetic disorders of metabolism" to being completely invisible in the conversation, even though they're where methylation actually matters most in a causative sense.
They're in neonatal intensive care units and pediatric neurology clinics. The wellness conversation has essentially captured a term that properly belongs to a set of rare, devastating genetic diseases and repurposed it for a much milder, much more common set of concerns.
Which is not to say that methylation isn't relevant to common diseases. The cancer connection is real — aberrant methylation of tumor suppressors is a hallmark of many malignancies. The aging connection is real. But those are patterns of dysregulation, not "low methylation" that you can fix with a pill.
And the direction matters. In some cancers you want to demethylate silenced tumor suppressors. In others you want to methylate activated oncogenes. There's no universal direction of "good" methylation. Context is everything.
Let's talk about the epigenetic editing angle a bit more, because I think that's where this is heading. If you can target methylation at specific genes without cutting DNA, you've essentially decoupled the therapeutic effect from the permanent genetic risk. But that also means you're treating a downstream effect, not the root cause. In an IEM, the broken enzyme is still broken. You're just managing the consequences more elegantly.
That's a fair critique. Epigenetic editing for IEMs would be a sophisticated form of management, not a cure. The cure would be gene therapy that delivers a working copy of the defective enzyme, or gene editing that corrects the mutation itself. But those approaches have their own challenges — immune responses to the viral vector, off-target editing, and in the case of gene therapy, the fact that the therapeutic gene often isn't regulated properly because it's not in its natural genomic context.
Epigenetic editing might actually be more physiologic in some ways. You're coaxing the patient's own gene to work harder rather than introducing an artificial copy.
But it's also less durable. If the epigenetic mark fades over time, you need repeat treatments. A one-time gene therapy that permanently fixes the defect is more appealing from a convenience standpoint, assuming it's safe.
Trade-offs everywhere.
Now: Hilbert's daily fun fact.
The average cumulus cloud weighs approximately one point one million pounds, roughly the same as eighty elephants.
For listeners who want something practical out of this — and Daniel's prompt was partly about understanding what's real versus what's noise — what should someone do if they're concerned about methylation, either for themselves or for a family member?
First, understand what you're actually worried about. If you have vague symptoms like fatigue or brain fog, and someone has told you that you have an MTHFR mutation and need methylated supplements, I would be skeptical. The common MTHFR variants are not disease-causing. The symptoms those supplements claim to treat are nonspecific and have dozens of more likely explanations — sleep deprivation, iron deficiency, thyroid dysfunction, depression. Chase those first.
If there's a family history of a metabolic disorder, or a child with unexplained neurologic symptoms?
That's a different scenario entirely. Newborn screening in most developed countries now covers many IEMs, including some remethylation disorders. But if you have a child who was born before expanded screening was implemented, or who has symptoms that emerged later, a metabolic geneticist is the right specialist. Plasma amino acids, plasma homocysteine, urine organic acids — these are first-line tests. They're not exotic. Any major children's hospital can run them.
For adults with unexplained neurologic issues — early stroke, progressive cognitive decline, neuropathy — homocysteine is the test to ask about. It's cheap, it's widely available, and if it's elevated, it opens up a whole diagnostic pathway that might otherwise be missed.
If it is elevated, the workup involves distinguishing between nutritional B twelve or folate deficiency, which is common and easily treated, and an actual remethylation disorder, which is rare and requires specialized management. A metabolic geneticist can make that distinction.
The other practical takeaway is just intellectual hygiene around health claims. When someone says "methylation is the root cause of disease," ask: which disease? Hypermethylated or hypomethylated? What's the evidence that changing methylation changes outcomes? Most of the time, the answers aren't there. The science is real, but it's been flattened into a marketing narrative that bears little resemblance to the underlying biology.
I'd add: if you're interested in this area, the Rutten, Rots, and Oosterveer paper from twenty twenty is readable for a motivated non-specialist. It's open access, published in a major journal, and it lays out the metabolic-epigenetic connection without oversimplifying. That's a much better starting point than most of what you'll find on supplement websites.
One forward-looking thought before we wrap. The convergence of metabolomics and epigenomics is going to reshape how we think about a lot of diseases, not just IEMs. The idea that your metabolic state directly modifies your gene expression through methylation and other epigenetic marks — that's a framework that applies to diabetes, to obesity, to cancer, to aging. We're just starting to map those connections systematically. In ten years, I suspect "methylation" will be a much more precise clinical term and a much less useful marketing term.
I hope so. The tools are getting better. Whole-genome methylation sequencing is becoming cheaper. Metabolomics is becoming more accessible. The resolution is improving. We're moving from "your methylation is off" to "this specific CpG site in the promoter of this specific gene is hypermethylated in this specific tissue, and here's the metabolite driving it." That's actionable in a way that the current vague narratives are not.
Thanks to our producer Hilbert Flumingtop for making this episode happen. This has been My Weird Prompts. You can find every episode at myweirdprompts dot com.
If you enjoyed this, leave us a review — it helps other people find the show. We'll be back soon with more from Daniel's ever-growing list of questions.
