Slippery definitions: how neo-darwinism accommodates contrary findings

There is a long history to this. There have been many points at which neo-darwinism has been challenged experimentally. The reactions can be characterised as a mixture of assimilation and denial.

An early example of assimilation is the reaction to the work of Conrad Waddington who first coined the term ‘epigenetics’. Waddington showed that an acquired characteristic in fruit flies could be assimilated after a number of generations of selection for the trait. After that it became inheritable in the standard way without the environmental stimulus that caused the character to be acquired. In the Experimental Physiology article I write

The Modern Synthesists should not have dismissed Waddington’s experiments, for example, as simply ‘a special case of the evolution of phenotypic plasticity’ (Arthur, 2010). Of course, the Modern Synthesis can account for the inheritance of the potential for plasticity, but what it cannot allow is the inheritance of a specific acquired form of that plasticity. Waddington’s experiments demonstrate precisely inheritance of specific forms of acquired characteristics, as he claimed himself in the title of his paper (Waddington, 1942). After all, the pattern of the genome is as much inherited as its individual components, and those patterns can be determined by the environment.

Notice that this also illustrates how neo-darwinism tends to be slippery when it defines a gene. Defining genes as individual protein-coding sequences ignores inheritance of characteristics that depend on the pattern of the genome. Waddington’s experiments show that it is perfectly plausible that a characteristic first arises through environmentally-induced change and then becomes ‘locked in’ to the genome pattern so that it becomes standard genomic inheritance. No mutations would be required and the characteristic could appear in a substantial fraction of the population rather than a single individual. We don’t know how often this process may have occurred during evolution. I suspect that many of the additional mechanisms of variation may become ‘locked in’ in this way.

A ‘gene’ in this sense can be a new combination of alleles that allows a new phenotype to develop. The mistake is to ‘atomise’ the concept of a gene. That is another important difference between the original definition in terms of phenotypes and the molecular biological definition in terms of ‘atomised’ DNA elements.

An example of what I would call denial is provided by an exchange during the 2009 Oxford debate that I chaired between Margulis and Dawkins. The transcript includes Dawkins’ reaction to the example of cellular inheritance independent of the genome in Paramecium:

"One example you might have meant is Sonneborn’s Paramecium where you cut a bit of the pellicle [ciliate cortex] and twist it ‘round. Well, if that’s true, and is indeed a non-DNA form of heredity, that’s absolutely fine. I would embrace that gladly as a new “honorary” gene. That’s fine. [Groans from the audience] Why not, why not?" (transcript, page 21)

The groans from the audience surely indicate something slippery here. I’d call it 'having it both ways'. The whole point of the discussion was whether inheritance is determined by genes alone. Sure, we can go on redefining 'gene' until we have exhausted all the possibilities. But a statement that says everything also says nothing. The denial here is a denial of a counter-example to neo-darwinism by redefining what is meant by a gene. As I explain in the lectures the central confusion lies precisely in the slippage between defining genes as DNA sequences and their original definition in terms of inheritable phenotypes. Defined as the latter, there can, by definition, be no exceptions. Defined as the former, there clearly are exceptions.

The relevant slide in the lecture is:

Relations between genes, environment and phenotype characters according to current physiological and biochemical understanding This diagram represents the interaction between genes (DNA sequences), environment and phenotype as occurring through biological networks. The causation occurs in both directions between all three influences on the networks. This view is very different from the idea that genes ‘cause’ the phenotype (right hand arrow). This diagram also helps to explain the difference between the original concept of a gene as the cause of a particular phenotype and the modern definition as a DNA sequence. For further description and analysis of the ideas behind this diagram see Kohl et al. (2010) Clinical Pharmacology and Therapeutics 88, 25–33 .

The sense in which neo-darwinism has been falsified in this case is precisely that epigenetic inheritance shows that characteristics of the biological networks can be inherited independently of DNA. Examples of that are now rapidly accumulating.

Are these kinds of confusions still occurring? A recent example is provided by

Zuk et al (2012) The mystery of missing heritability: Genetic interactions create phantom heritability. PNAS, 109, 1193-1198.

which contains an equivalent to Dawkins’ invention of ‘honorary gene’ by inventing the term ‘phantom heritability’. In this case, the interactions that necessarily include the biological networks are apparently assimilated into standard genetics (hence the term ‘genetic interactions’), when in fact there must be much more inherited than the relevant DNA sequences. The authors write “In short, missing heritability need not directly correspond to missing variants, because current estimates of total heritability may be significantly inflated by genetic interactions.” The matter should not be closed off in this way since the question whether this does or does not involve inheritance of epigenetic effects must remain open to further experimentation. And it would be better to recognise this by avoiding terms like ‘genetic interactions’ that leave it unclear what precisely is involved but which also prejudice the conclusion in favour of standard genetic inheritance.

Redefinition of the central entity in neo-darwinism, i.e. the ‘gene’, leads to unintended consequences. Admitting ‘honorary’ or ‘phantom’ genes that refer entirely or partly to cytoplasmic inheritance makes nonsense of the distinction between ‘replicator’ and ‘vehicle’. The ‘vehicle’ becomes part of the ‘replicator’.

It is anyway! See Immortal genes.