More Recent Comments

Saturday, April 03, 2021

"Dark matter" as an argument against junk DNA

Opponents of junk DNA have been largely unsuccessful in demonstrating that most of our genome is functional. Many of them are vaguely aware of the fact that "no function" (i.e. junk) is the default hypothesis and the onus is on them to come up with evidence of function. In order to shift, or obfuscate, this burden of proof they have increasingly begun to talk about the "dark matter" of the genome. The idea is to pretend that most of the genome is a complete mystery so that you can't say for certain whether it is junk or functional.

One of the more recent attempts appears in the "Journal Club" section of Nature Reviews Genetics. It focuses on repetitive DNA.

Before looking at that article, let's begin by summarizing what we already know about repetitive DNA. It includes highly repetitive DNA consisting of mutliple tandem repeats of short sequences such as ATATATATAT... or CGACGACGACGA ... or even longer repeats. Much of this is located in centromeric regions of the chromosome and I estimate that functional highly repetitve regions make up about 1% of the genome.[see Centromere DNA and Telomeres]

The other part of repetitive DNA is middle repetitive DNA, which is largely composed of transposons and endogenous viruses, although it includes ribosomal RNA genes and origins of replication. Most of these sequences are dispersed as single copies throughout the genome. It's difficult to determine exactly how much of the genome consists of these middle repetitive sequences but it's certainly more than 50%.

Almost all of the transposon- and virus-related sequences are defective copies of once active transposons and viruses. Most of them are just fragments of the originals. They are evolving at the neutral rate so they look like junk and they behave like junk.1 That's not selfish DNA because is doesn't transpose and it's not "dark matter." These fragments have all the characterstics of nonfunctional junk in our genome.

We know that the C-value paradox is mostly explained by differing amounts of repetitive DNA in different genomes and this is consistent with the idea that they are junk. We know that less that 10% of our genome is conserved and this fits in with that conclusion. Finally, we know that genetic load arguments indicate that most our genome must be impervious to mutation. Combined, these are all powerful bits of evidence and logic in favor of repetitive sequences being mostly junk DNA.

Now let's look at what Neil Gemmell says in this article.

Gemmell, N.J. (2021) Repetitive DNA: genomic dark matter matters. Nature Reviews Genetics:1-1. [doi: 10.1038/s41576-021-00354-8]

"Repetitive DNA sequences were found in hundreds of thousands, and sometimes millions, of copies in the genomes of most eukaryotes. while widespread and evolutionarily conserved, the function of these repeats was unknown. Provocatively, Britten and Kohne concluded 'a concept that is repugnant to us is that about half of the DNA of higher organisms is trivial or permanently inert.'”"

That's from Britten and Kohne (1968) and it's true that more than 50 years ago those workers didn't like the idea of junk DNA. Britten argued that most of this repetitive DNA was likely to be involved in regulation. Gemmell goes on to describe centromeres and telomeres and mentions that most repetitive DNA was thought to be junk.

"... the idea that much of the genome is junk, maintained and perpetuated by random chance, seemed as broadly unsatisfactory to me as it had to the original authors. Enthralled by the mystery of why half our genome is repetitive DNA, I have followed this field ever since."

Gemmell is not alone. In spite of all the evidence for junk DNA, the majority of scientists don't like the fact that most of our genome is junk. Here's how he justifies his continued skepticism.

"But it was not until the 2000s, as full eukaryotic genome sequences emerged, that we discovered that the repetitive non-coding regions of our genome harbour large numbers of promoters, enhancers, transcription factor binding sites and regulatory RNAs that control gene expression. More recently, the importance of repetitive DNA in both structural and regulatory processes has emerged, but much remains to be discovered and understood. It is time to shine further light on this genomic dark matter."

This appears to be the ENCODE publicity campaign legacy rearing its ugly head once more. Most Sandwalk readers know that the presence of transcription factor binding sites, RNA polymerase binding sites, and junk RNA is exactly what one would predict from a genome full of defective transposons. Most of us know that a big fat sloppy genome is bound to contain millions of spurious binding sites for transcription factors so this says nothing about function.

Apparently Gemmell's skepticism doesn't apply to the ENCODE results so he still thinks that all those bits and pieces of transposons are mysterious bits of dark matter that could be several billion base pairs of functional DNA. I don't know what he imagines they could be doing.

Photo Credit: The photo shows human chromosomes labelled with a telomere probe (yellow), from Christoher Counter at Duke University.

1. In my book, I cover this in a section called "If it walks like a duck ..." It's a form of abductive reasoning.

Britten, R. and Kohne, D. (1968) Repeated Sequences in DNA. Science 161:529-540. [doi: 10.1126/science.161.3841.529]


  1. Hi dr.larry
    I am an begginer in Genomics

    I wanna ask you about pseudogenes

    how do we identify any sequence in DNA as a pseudogene?
    I know that it shouldn't code for a protein and must resemble another sequence that codes for a protein
    But are there more strong mechanisms for identifying a pseudogene?
    Especially unitary pseudogenes.

    1. Is it disabled by a frame-shift indel or a stop codon? And is it orthologous to a function gene in a related species? That would be a start.

    2. No,
      When I give creationsts some evidence for evolution by using pseudogenes they say:
      1-ok,your argument will be strong if these sequences wich looks like the original gene have no functions

      if they have an functions (not necessary encode protein) we can assume that creater has created these genes and they are not genes disabled by mutations
      it just the creator have created them look like other genes wich encode protein.

      so I wanna know:
      Are there any methods to identify pseudogenes (especially unitary pseudogenes) other than compare the sequences?

    3. John is correct for protein-coding genes but it's more difficult to identify a pseudogene derived from a noncoding gene. Nevertheless, there are still plenty of those and they're recognized because they are evolving at the neutral rate.

      The very first pseudogene ever discovered was a 5S RNA pseudogene in Xenopus (1977) although you could argue that the ABO pseudogene (a polymorphic pseudogene) was inferred long before that.

    4. it just the creator have created them look like other genes wich encode protein.

      Ask them why the creator would do that. Why create a sequence that looks like a protein-coding sequence in some other species, in the same place in the genome as the protein-coding gene is in those other species? And as Larry said, why does it appear to evolve at the neutral rate? GULO in primates is the most frequent example.

      And of course you can easily assume that any sequence has an unknown function, and it's impossible to prove otherwise. You could however show that it did have a function, and a few pseudogenes do, though not GULO. Wouldn't the burden in such cases be on the creationist to show that the pseudogene is actually functional? Still, isn't a functional pseudogene still a pseudogene, as long as it doesn't have the function of the original gene?

  2. Hi Larry,I just discovered your blog and really like your posts here especially the ones on junk DNA. I never knew that almost half the genome is composed of transposons and have alway bought the line that most of our genome is functional but just waiting to be discovered.

    From my brief reading however, it seems that transposons can have an effect on the expression of nearby genes and are involved in various diseases. So it may certainly still be 'noncoding' but not necessarily non-functional. Do correct me if I misread your post.

    1. Just because some small number of transposons have secondarily acquired a function does not mean that ALL transposon-related sequences in our genome have a function. The null explanation has to be junk so the onus is on the opponents of junk DNA to show that most of our genome is functional.

      Also, please don't use noncoding DNA as a synonym for junk DNA.