Really? While we all can agree that there's no definition of "gene" that doesn't have exceptions, we can surely agree that some definitions work pretty well. I've argued that defining a gene as, "a DNA sequence that is transcribed to produce a functional product" works well in most cases [What Is a Gene?].
Let's see how James Shapiro handles this problem.1 He says,
The identification of DNA as the key molecule of heredity and Crick's Central Dogma of Molecule Biology [Crick 1970] initially seemed to confirm Beadle and Tatum's "one gene -- one enzyme" hypothesis.I've already explained that Shapiro doesn't understand the Central Dogma of Molecular Biology even though he quotes the Francis Crick papers that explain it correctly [Revisiting the Central Dogma in the 21st Century]. I also made this point in my review of his book: Evolution: A View from the 21st Century.
The Central Dogma says that "... once (sequential) information has passed into protein it cannot get out again." In other words, you can't transfer sequence information from protein to nucleic acid [Basic Concepts: The Central Dogma of Molecular Biology]. This has nothing to do with the concept of a gene and it certainly doesn't confirm the old "one gene—one enzyme" hypothesis. At the time that Crick wrote his 1970 Nature paper—the one always referred to by Shapiro—he certainly knew about genes for ribosomal RNA and genes for tRNAs. These are clear examples of genes that don't encode proteins.
Shapiro has read my blog and he has read my review. Presumably he has by now (re)read Crick's 1970 paper. So how come he still doesn't get it?
Shapiro continues ...
However, molecular genetics quickly introduced difficulties with the theory of atomistic genes aligned like beads on a string.None of this was a challenge to our concept of "gene" because junk DNA and repetitive DNA isn't genes. Now, maybe if you believed that all genes encode proteins back in 1970 you might have been confused by all that, but most of us abandoned Beadle and Tatum almost fifty year ago. Apparently Shapiro is a slow learner.
A major challenge was Britten and Kohne's1968 discovery of massive amounts of repetitive DNA in certain genomes. Today, we know our DNA contains over 30 times as many base-pairs in repeats as it does in protein coding sequences. By the conventional view, if genes are the only important actors, then these surprisingly abundant "intergenic" repeats must constitute "junk DNA" and be "ultimate parasites" in the genome.
As readers of this and other science blogs know well, the junk DNA idea has been challenged by the large-scale ENCODE project, designed to produce the "Encyclopedia Of DNA Elements" independently of theoretical prejudices. In its first few years, ENCODE has documented cell type-specific biochemical activity in over 80 percent of this repetitive DNA and known functions in 20 percent.Oh dear! Most intelligent molecular biologists are well aware of the problems with the ENCODE claim. Shapiro isn't.
The basic issue is that molecular genetics has made it impossible to provide a consistent, or even useful, definition of the term "gene."
But even if ENCODE's most incredible claim were true it still wouldn't have any effect on our definition of "gene." If lots of our genome is transcribed into functional RNA molecules then it simply means that we have more genes than we originally thought. If lots of the genome consists of regulatory sequences and DNA binding sites then that's fine too. These aren't genes—they are sequence that regulate the expression of genes.
Since the 1970s it's been possible to think of a gene as a "DNA sequence that's transcribed" without inducing any brain seizures. Nothing has changed since then in spite of what James Shapiro would have you believe.
1. Jerry Coyne has already dissected Shapiro's article at: James Shapiro gets evolution wrong again.