There's no such thing as a Selfish Gene | Denis Noble on Dawkins' mistake
What this covers
Denis Noble, a systems biologist at Oxford, argues that gene-centric biology rests on four foundational errors that have inverted the actual causal relationship between organisms and their genetic material. Rather than genes controlling organisms, Noble contends that organisms actively control their genomes—a reversal with significant consequences for both theory and medical practice. The talk walks through each error in turn: the central dogma, the Weismann barrier, DNA self-replication, and the replicator-vehicle separation that underpins the selfish-gene model. Noble draws on molecular evidence, including a 2025 Nature article, to show how each of these concepts misrepresents what the evidence actually demonstrates about how cells manage their DNA.
A conceptual problem runs through the entire framework. DNA nucleotides are chemicals governed by binding energies—adenine pairs with thymine, guanine with cytosine—with no capacity for choice. Describing genes as "selfish" therefore commits a category error; only organisms with agency can exhibit selfishness or cooperation. Beyond the theoretical correction, Noble points to practical failure: genome sequencing has not delivered the promised medical breakthroughs. Polygenic scoring, tested by researchers at UCL on cardiovascular disease and cancer—conditions accounting for 60 percent of deaths—simply does not predict disease incidence. Meanwhile, the Weismann barrier, long thought to isolate germline cells from bodily influence, has been breached: vesicles carrying regulatory molecules cross from body cells into germline cells and alter their function. DNA's self-replication, presented in standard accounts as an automatic chemical process, actually depends entirely on a "controlled army" of cellular repair enzymes that reduce error rates from one in ten thousand to one in ten billion per division. These observations suggest organisms maintain robust backup systems for most genes, which explains why single-gene knockouts often produce modest effects.
Standard gene-centric biology has causation upside down: genes are passive chemicals that organisms actively control, so the central dogma, Weismann barrier, DNA self-replication, and the selfish-gene model are all incorrect as complete explanations — and this error explains why genome sequencing has failed to deliver cures for common complex diseases.
- DNA nucleotides are chemicals with no choice, so 'selfish' is a category error; only organisms that can choose are selfish or cooperative
- Four foundational dogmas (central dogma, Weismann barrier, DNA self-replication, replicator-vehicle separation) are refuted by molecular evidence in a 2025 Nature article
- Polygenic scores fail to predict cardiovascular disease and cancer, showing gene-disease associations are generally weak and organisms back up most genes
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
All four dogmas (central dogma as complete story, Weismann barrier, DNA self-replication, replicator-vehicle separation) are shown incorrect by the correct interpretation of molecular biological evidence, as published in a Nature article in February of this year.
“the correct interpretation of the molecular biological evidence shows that all of these dogmas are incorrect. That's in the nature article published in February this year.”
Because DNA cannot self-replicate without the cell's repair machinery, the replicator cannot be represented as separate from its vehicle, which refutes the fundamental thesis of the selfish gene.
“The replicator cannot be represented as separate from its vehicle. That's the fundamental thesis of the selfish gene and is the reason why two years ago I said it's incorrect.”
DNA nucleotides are chemicals whose associations and dissociations are determined purely by binding energies (A binds T, G binds C) and have no choice, so describing genes as 'selfish' — metaphorically or literally — is incorrect because only organisms with the freedom to choose can be selfish or cooperative.
“as chemicals... they can only do what chemicals naturally do. Their associations and dissociations from other chemicals in the body are determined by the energies of those associations or disassociations... And in doing this they have no choice. And that is the fundamental reason why it's incorrect to describe them as selfish”
Standard biology has causation upside down: genes do not control organisms ('genes create us body and mind' per Dawkins); rather, organisms control and change their genes when needed.
“Standard biology, I'm sad to say because I'm a biologist myself... has got causation upside down. Genes, to quote Richard Dawkins, genes create us body and mind. I reply, we control them and change them when we need to.”
DNA cannot self-replicate like a crystal (an idea tracing to Schrödinger's 1942 Dublin lectures): chemical base-pairing alone yields an error rate of about 1 in 10,000 base pairs, which across 3 billion base pairs would produce hundreds of thousands of errors per division; the living cell instead deploys a controlled army of cut-and-paste correction enzymes that lower the error rate to about 1 in 10 billion.
“that process of self-replication... is accurate up to a point... but like all chemical reactions it has an energy of formation and disassociation... It's one in 10,000 base pairs... We have three billion base pairs. There would be hundreds of thousands of errors with that process... The cell, the living cell itself has an army of cut and paste enzymes which it controls to achieve the correction... You end up with an error rate of only about 1 in 10 billion”
The Weismann barrier (Weismann, 1883), the claim that germline cells are separated from the body such that bodily changes can never be communicated to the germline, is wrong: all body cells extrude tiny vesicles containing control molecules (RNAs, proteins) that cross the barrier and have been shown to affect the metabolism and physiology of germline cells.
“the viceman barrier which is that... there's a barrier between the two so that the body changes can never be communicated to the germ line... we've now found that all cells in the body extrude tiny packets... Many of those bubble over and pass across the so-called barrier and have now been shown to affect the metabolism and other major physiological properties of the germline cells.”
Genome sequencing has largely failed to deliver promised medical treatments for common complex diseases: a UCL team published in the BMJ in October last year showed that polygenic scores, built using the same techniques as a clinical drug trial, simply do not work for cardiovascular disease and cancer — which together account for 60% of fatalities — though monogenic (rare single-gene) diseases remain a valid target for genetic cures.
“University College London, a big team, put that idea to the test... published in October last year in the British Medical Journal... form what is called a polygenic score... it simply doesn't work for cardiovascular disease and cancer. those two account for 60% of fatalities.”
Francis Collins promised in the 1990s/2001 that within 10 years human genome sequencing would yield 'previously unimaginable insights' and 'rational strategies for minimizing and preventing the disease phenotypes altogether.'
“He wrote that within 10 years of human genome sequencing, it would lead to, and I quote, exactly, previously unimaginable insights and from there to the common good, including... the development of rational strategies for minimizing and preventing the disease phenotypes altogether.”
The central dogma of molecular biology (Crick, 1958), that information flows DNA→RNA→protein and cannot be reversed, is correct as a chemical sequence statement, but it does not prevent organisms from controlling their genomes — as shown when immune systems mutate only the antigen-grabbing region of immunoglobulin genes via chance changes, then select and proliferate the cells that successfully grab the pathogen and kill the rest.
“the central dogma of molecular biology formulated by Crick in 1958... absolutely correct... But that is not the way in which organisms control their genomes. What was happening during the pandemic illustrates this perfectly well... They then proceeded to mutate the relevant part of the genome and only that part that codes for the grabbing part of the protein”
Association scores between the presence/absence of most genes and the incidence of major disease are generally very low; in the speaker's own Oxford work on the heart's pacemaker mechanism, knocking out a causally important gene produced only a very modest frequency change, demonstrating that organisms back up the great majority of their genes.
“I showed that in my own work with a team of scientists in Oxford looking at the pacemaker mechanism in the heart... you can knock a major gene out that we know is causally important... and the frequency change is very modest. So it's clear that organisms back up the great majority of the genes”
The human genome is a thread of roughly three billion nucleotides (A, T, G, C), and every cell except red blood cells contains a complete set.
“in humans... there are at least three billion nucleotides... All ourselves except for the red blood cells contain a complete set of those.”
A baby is not born selfish; it simply has needs (food and care) to live, grow and flourish, and over about three or four years it learns that it can choose.
“A baby, therefore, is not born selfish. It simply has needs, food, and care... And it slowly learns that it can choose. And after about three or four years... it can choose and it knows what it wants but for those few years it doesn't.”