For years, scientists have known two things about autism spectrum disorder (ASD):

  1. It’s highly heritable.
  2. Kids with ASD often have different gut bacteria than neurotypical kids.

But here’s the twist:

It may not just be which bacteria you have.

It may be which versions of those bacteria you have.

And that’s a much deeper level of biology.

First, The Basics

ASD is a neurodevelopmental condition marked by:

  • Social communication differences
  • Restricted or repetitive behaviors
  • Sensory sensitivities

Its prevalence has risen dramatically in recent decades, and while genetics plays a major role, it doesn’t explain everything.

That’s where the gut comes in.

Researchers have repeatedly found that children with ASD have different gut microbiomes compared with neurotypical peers — including shifts in bacteria like Bacteroides and Ruminococcus.

But most studies stop at abundance:

“How much of bacteria X is present?”

This new research went further.

Much further.

The Big Idea: Bacteria Have Their Own Genetic Variations

Just like humans have genetic differences, bacteria do too.

Inside a single bacterial species, there can be:

  • Gene gains
  • Gene losses
  • DNA rearrangements

These are called structural variations (SVs).

And they can change how bacteria behave — including how they:

  • Process nutrients
  • Produce metabolites
  • Interact with the immune system
  • Influence the gut-brain axis

Until now, no one had systematically studied bacterial structural variations in autism.

So this team did.

The Study: 452 Kids, Global Data

Researchers analyzed stool samples from:

  • 261 children with ASD
  • 191 neurotypical children

Using shotgun metagenomic sequencing (which reads bacterial DNA in detail), they looked not just at species — but at genomic differences within species.

What they found?

Significant bacterial genomic differences linked to ASD.

The Two Bacteria That Stood Out

1) Bacteroides uniformis

This species is usually considered a beneficial gut bacterium.

But in children with ASD, researchers found:

  • A missing genomic region linked to thiamine (vitamin B1) and iron metabolism
  • Another variation affecting exopolysaccharide (EPS) production — important for bacterial adhesion and gut colonization

Why does that matter?

Because:

  • Thiamine and iron are critical for brain development
  • EPS affects how bacteria stick to the gut lining
  • Gut barrier function influences inflammation and the gut-brain axis

In mouse experiments, the thiamine/iron-related variation was associated with:

  • Altered social interaction
  • Increased repetitive behaviors

That’s a major signal.

2) Ruminococcus torques

This bacterium was found in higher abundance in ASD.

Researchers discovered:

  • A depleted genomic region encoding MazF — a bacterial “growth brake” protein
  • Another region affecting α-galactosidase, an enzyme involved in breaking down mucus glycans

Translation:

If growth-inhibiting genes are missing, the bacteria may overgrow.

If mucus-processing genes are altered, the gut barrier may weaken.

That could mean:

  • Increased gut permeability
  • Altered immune signaling
  • Disrupted gut-brain communication

This aligns with long-standing theories about gut barrier dysfunction in ASD.

The Bigger Pattern: Metabolic Disruption

Across multiple bacteria, the study found genomic variations tied to:

  • Amino acid metabolism
  • Lipid metabolism
  • Iron regulation
  • Carbohydrate processing

These pathways connect directly to:

  • Neurotransmitter production
  • Inflammation
  • Energy metabolism

In other words:

The bacterial genome changes matched known metabolic irregularities observed in ASD.

That’s not coincidence. That’s alignment.

Geography Matters

The researchers also found something fascinating:

Different populations had different microbial genomic signatures.

Western cohorts showed variations tied to mucosal adhesion — possibly reflecting lower-fiber diets.

Eastern cohorts showed variants linked to carbohydrate metabolism and stress adaptation.

Translation?

Diet and environment shape bacterial genetics.

And those differences may interact with ASD biology.

This means future diagnostics may need to be region-specific.

The Diagnostic Breakthrough

Here’s where things get practical.

The team built a diagnostic model combining:

  • Bacterial structural variations
  • Bacterial abundance data

The result?

An AUROC of:

  • 85.8% in the discovery group
  • 81.1% in the validation group

For context:

Many blood-based ASD biomarker studies show wide performance variability.

This method is:

  • Non-invasive (stool sample)
  • Child-friendly
  • Suitable for repeated monitoring

And importantly:

More precise than using bacterial abundance alone.

Because instead of asking “Is this species present?”
It asks, “Which functional version of this species is present?”

That’s a much sharper tool.

What This Doesn’t Mean

Let’s pump the brakes.

This study was cross-sectional.

That means:

It shows association.
Not causation.

We don’t yet know:

  • Whether bacterial genomic changes cause ASD
  • Or whether ASD-related physiology shapes bacterial genomes

It also didn’t include host genetic sequencing.

Given ASD’s strong genetic basis, future research must combine:

  • Human genomics
  • Microbial genomics
  • Longitudinal data

Why This Matters

For years, the gut-brain axis conversation has been high-level:

“Autism and the microbiome are linked.”

This study drills deeper.

It suggests:

Not just microbial imbalance.

But microbial genetic rewiring.

And that rewiring may affect:

  • Nutrient metabolism
  • Immune signaling
  • Gut barrier integrity
  • Behavioral outcomes

That’s a mechanistic foothold.

The Summary

Autism is complex. Deeply complex.

Genes matter.
Environment matters.
Biology is layered.

This study adds a new layer:

The genetic variation inside gut bacteria.

And that variation may shape how microbes talk to the brain.

The practical upside?

A stool-based diagnostic panel that could:

  • Improve early detection
  • Help stratify subtypes
  • Guide microbiome-targeted therapies

The scientific upside?

We’re moving from:

“Which bacteria are there?”

To:

“What are those bacteria actually capable of doing?”

And that shift — from presence to function — could be a turning point in microbiome research.

The gut-brain axis just got a lot more genomic.

Liu W, Lu Y, Ng SC, Chan FK, Sung JJ, Yu J. Bacterial genomic structural variations in children with autism serve as diagnostic biomarkers. Gut. 2026 Feb 26:gutjnl-2025-337280. doi: 10.1136/gutjnl-2025-337280. Epub ahead of print. PMID: 41748159.