Series: Chasms of Evolutionary Impossibilities – Douglas Axe’s Work (2004) and the Evolutionary Impossibility of a Mere Protein.

doi:10.1016/j.jmb.2004.06.058

8.2 “Axe Ignored the Neutrality of Mutations”

When theory seems elegant — but data shows it doesn't work

Objection

Some critics say that Douglas Axe ignored an important aspect of evolution: the existence of neutral mutations. These would be genetic changes that do not harm protein function and could therefore accumulate over time, serving as "bridges" for new evolutionary functions to appear gradually.

🪜 For the lay reader: It is like saying you can change some parts of a car without affecting its operation — and that, over time, these changes could turn the car into an airplane. It seems like an ingenious idea, but does it work in practice?

What Axe Actually Did

Axe did not ignore this hypothesis — he tested it directly. He took a functional protein (β-lactamase, which helps bacteria resist antibiotics) and created 10⁷⁷ different versions of it, each with small alterations. The goal was to see if any of these versions still worked.

✅ Accessible explanation:

• Imagine the protein is a complex lock
• The modified versions are slightly different keys
• Axe wanted to know: do these keys still open the lock?

🪜 Integrated analogy:

“It is like trying to open a door with a key that has been filed just a little. Even small changes can make it stop working.”

Axe was even more generous: he tested these variants at lower temperature (25 °C), which facilitates the activity of unstable proteins. It is like testing a car on a flat track, with ideal weather, to see if it works. Even so, no neutral variant with significant function was found.

Where is the Logical Error?

The criticism assumes that neutral mutations are common and functional — but that is precisely what needs to be proven. Axe tested this idea and showed that, in practice, most mutations are not neutral — they are harmful.

🪜 Explanation for laypeople: It is like saying you can change random letters in a computer program's code and expect it to keep working. Most of the time, the program crashes or gives an error. And rarely, when it doesn't crash, it does nothing new.

What the Data Show

The results from Table 3 of the study show that even minimal changes reduce the protein's efficiency by 100 to 10,000 times. This means the protein's function practically disappears with subtle alterations.

$$P(\text{function}) \approx 10^{-150}$$

✅ Didactic explanation: Imagine a functional protein is like a sequence of 150 phone numbers in exact order. The chance of guessing this sequence by random trial is less than 1 in 10¹⁵⁰.

Now, neutrality advocates suggest we can miss a few numbers without problem. But the data show that missing even a single number makes the chance of success plummet. If we need 5 beneficial mutations simultaneously, the probability becomes:

$$P \approx 10^{-600}$$

🪜 For the lay reader: This number is so small that even if all atoms in the universe were computers testing combinations for billions of years, we couldn't find a functional sequence by chance.

Model

Axe used biochemical measures to evaluate whether the modified proteins still functioned:

• ΔG (stability): shows if the protein maintains its shape — like a well-pitched tent
• k_cat/K_M (efficiency): indicates if the protein performs its function — like an engine that runs with power
• Test temperature: was adjusted to favor unstable variants — like testing a car on a flat track to see if it runs

🪜 Functional analogy:

“He didn't just test if the car started — he tested if it ran well, with ideal fuel, under perfect conditions. And even then, the altered models failed.”

Fixation Time and Genetic Entropy

Even if a neutral mutation occurs, the time needed for it to fix in a population is given by:

$$t_{\text{fix}} = 4N_e$$

For bacteria with \(N_e \approx 10^9\), this represents about 4 billion generations. With 100 generations per year, it would take 40 million years to fix a single neutral mutation.

🔧 Genetic entropy problem: Even neutral mutations accumulate genetic deterioration. Sanford (2005) demonstrates:

$$\Delta H = -k \sum p \ln(p) > 0$$

Where genetic entropy \(\Delta H\) increases inexorably in real populations, preventing the preservation of emerging functions.

Source: Sanford (2005) – Genetic Entropy & The Mystery of the Genome

🪜 Explanation for laypeople: It is like trying to keep a text readable while random letters are being changed over time. Even if some changes don't harm immediately, the inevitable accumulation leads to loss of meaning.

What Does the Scientific Literature Say?

• Lynch (2020): Admits that "most non-synonymous mutations are deleterious even under permissive conditions"
• Koonin (2016): Recognizes that "the space of neutral sequences is extraordinarily restricted for functional proteins"
• Tokuriki & Tawfik (2009): Show there is a trade-off between flexibility and function — the more flexible, the less functional
• Axe (2004): Demonstrated experimentally that minimal mutations drastically compromise function

🪜 For the lay reader: Even scientists who defend evolution recognize that the margin for neutral mutations is much smaller than thought. The idea that proteins can change gradually without losing function is not confirmed by real data.

Why This Criticism Fails

The criticism fails because it presents no data contradicting Axe's results. It is based on a theoretical idea — that neutral mutations could accumulate — but this idea is not confirmed when tested experimentally.

🪜 Final analogy:

“It is like trying to improve a computer program by changing random letters in the code. Most of the time, the program breaks. And when it doesn't break, it doesn't improve.”

Conclusion for the Lay Reader

Douglas Axe did not ignore the neutrality of mutations — he tested it directly, with one of the most comprehensive experiments ever done on the topic.

And the data showed that even small alterations compromise the function of complex proteins.

The idea that neutral mutations can accumulate and generate new functions does not hold up against biochemical reality.

Real science, with measurable data, shows that the "neutrality space" is too small to allow random processes to build the complexity observed in nature.

🪜 Visual summary:

“Modifying functional proteins is like tinkering with a Swiss watch with a sledgehammer — any change tends to break the system.”

Therefore, this criticism does not invalidate the study.

It reinforces Axe's conclusion: functional evolution by random mutations is statistically improbable.

Priority Self-Refuting Sources (κ > 0.9)

• Lynch (2020): Admits fundamental limitations of evolutionary mechanisms
• Koonin (2016): Recognizes severe constraints in molecular evolution
• Axe (2004): Raw experimental data from Tables 1–3
• Tokuriki & Tawfik (2009): Confirms trade-off between flexibility and function
• Sanford (2005): Demonstrates cumulative genetic entropy in real populations