Michael, I agree with your statement of the definition of a mutation, at least in theme: I would say "a mutation is an event which changes the sequence of bases along one or more DNA molecules." There is nothing about mutation that restricts it to the gene level; likewise, there are several DNA damage repair pathways which correct mutations--so a mutation doesn't refer to just the permanent changes, but to evanescent ones as well.

I also think it is good to ask the kinds of questions you asked to clarify the issue. So, is a SNP (single-nucleotide polymorphism) a mutation? I think not. A SNP is, however, the _result_ of a mutation. The mutation is the change to the DNA molecule, which happened long ago and to a single molecule; the SNP is either present or not across a whole population of DNA molecules.

Is a chromosomal rearrangement a mutation? Yes because it is an event that causes the sequence of bases along (at least one) DNA molecule to change.

Is an error in DNA mismatch repair a mutation? Not necessarily, it depends on the error. If the error is that the mismatch repair system fails to restore the original sequence, then the error is NOT a mutation; the mutation already occurred. However, the DNA molecule is more likely to _transmit_ the mutation to progeny of the cell in this case. Technically speaking, _successful_ DNA mismatch repair is a mutation; it changes the DNA sequence back to what it was before the first mutation, which from a long-term view erases a mutation, however it fits both your definition and mine.

Regarding cross-species hybridization, I would say that even normal replication within a sexual species "leads to" mutation, as recombination events are very likely to be mutations by the definitions you and I both gave. In the same way, cross-species hybridization will generally involve recombination, and therefore mutation.

I also agree that there is a difference between using terminology formally, within a scientific discussion, and informally, trying to explain science without the level of precision required for experimentation. For instance, I have no problem with calling a SNP a mutation, when speaking to a sufficiently broad audience. In a research setting, though, being cavalier about what terms one uses can often lead to mis-stated hypotheses that don't actually correspond to any real, testable system. In my experience, it is often very helpful to consider exactly which term one should use to define an experiment; sometimes just this exercise alone can provide an answer to the original question, while in other cases it can show that a particular line of experimentation won't answer the question, and should be abandoned.

Emilio,

I think the definition I gave above is quite similar to "any change in DNA occurring in any time to a cell". My perspective is that this definition isn't "so ambiguous that it has no meaning", as you put it; it's just limited in use, as all terms are.

I think your frustration with these definitions stems from the difference between how we speak outside of research, where precise terms are often not necessary and can even detract from communication, and how detailed we need to be in our terms within a research setting. Terminology simply ISN'T context-dependent, no matter how much we might like it to be.

I should also say that I think we should always, always consider whether a term we use is well-defined with respect to the use we are making of it. For instance, do we need the term "two nucleotide polymorphism" or TNP to represent the consequences of a mutation that changed two consecutive nucleotides in a single event, or should we call that two adjacent SNPs? I can't answer that question in general, it depends on what use we plan on making of the terms SNP and TNP. In a context where we're only talking about one kind of mutation (single base substitutions, say) and only about allele frequencies within a population (as opposed to mutational mechanisms, for example), I think it makes discussions much clearer and briefer to call SNPs simply "mutations", as we've already narrowed the context to a setting where the two terms are essentially indistinguishable, and an acronym like SNP requires you to remember the three words it represents ("stable nuclear proton"?), rather than a clearer single word. On the other hand, if we're contrasting the mutational mechanisms that might change one base, then in a separate event change a neighboring base (two SNP mechanisms) versus a single mutational event that by its nature always changes two consecutive bases (one TNP), I think TNP is a justifiably useful term.

I disagree with you Jon. Not only in the presumptuous and offensive tone of yuur commentary but also in the content.

You say:

1: "a mutation is a change to genetic material in an organism (as opposed to nucleic acids in no role relevant to inheritance or information-specific contexts)"

But this is not correct because DNA methylation is a change in DNA and I never sought it considered as a mutation. In addition you have no apriori idea of when a piece of DNA may have a relevant role in inheritance.

Instead of your definition, i prefer to define a mutation as a change in the DNA sequence. More short, more precise.

2. "It is all very well wanting a clear, unambiguous term for every possible concept, with clear and unambiguous applications for each and every one, but language is small "

Language is not that small. There are many types of mutation and it is important to specify what we are talking about. In this case you don't need billions of words, just precisely define the type of mutation you refeer to.

DNA methylation isn't considered a mutation largely because it has generally been considered transitory, and until recently there were no good methods to study it at a large scale. With the newer technologies, that may change.

Thanks for your links Darryl. They assist clarifying gene mutation.

Any event that changes genetic structure; any alteration in the inherited nucleic acid sequence of the genotype of an organism)

Jon, I am glad to say that finally I agree with you . Please Darryl, delete that paragraph and I am sure Jon will vote up your answer this time.

Darryl - unfortunately you came to the wrong conclusion reading those web materials. A frameshift mutation is in fact a point mutation (generally an addition or deletion of one or two nucleotides) but resulting in a dramatic change to translation of a gene. You have to distinguish between the nature of the change itself (what base is changed to what, or deleted or added) versus what the effect of the change is to the phenotype of the organism. The term frameshift is only relevant in the context of translation of a gene, a single base insertion not within a reading frame is not a frameshift, but merely a single base insertion.

Daryl, the discussion as to whether frame shift mutations are point mutations or not, is entirely one of semantics at a level that is not germane to this discussion. There is an entire other class of events.

Geneticists would define mutations as a heritable change in function or phenotype. Not all changes in DNA sequence are mutations, they can be classified as variants or alleles, especially if they have no measurable phenotype. Indeed, most DNA substitutions or alterations have little to no impact on gene expression or function.

Historically, scientists discovered many mutations in mice and flies even before the structure of DNA or the nature of genetic material was known. So the definition of mutations predates the discovery of DNA.

Genetic mutation is any event that changes genetic structure; any alteration in the inherited nucleic acid sequence of the genotype of an organism.

I would like to define mutation in the context of my discipline that it's simply the change in base sequence of a test sample in comparison to a reference standard.