Genomes of prokaryotes tend to have a higher proportion of coding DNA to non-coding DNA in their sequence than eukaryotes. This means that either eukaryotes accumulated lots of non-coding DNA, much of which is regulatory, in the course of evolution — or that prokaryotes lost this proportion of non-coding DNA.

In order for transcription of a particular gene to occur, it seems that there must be some search performed on the DNA sequence to find the gene to be transcribed. This search seems to be independent of the sequence length — if it weren't, that could act as a selection pressure, inducing loss of DNA that doesn't do anything. Therefore, it seems that the method by which this search is done must be O(1) — that is, it takes a constant amount of time to complete the search, regardless of the length of the sequence. Such a method would be most welcome in computational sequence searches (in place of BLAST for example).

My question is, what, if anything, is currently known about the process of transcription that could explain this constant search time? Or, if I am completely wrong in my assumptions, what are more reasonable assumptions?

  • $\begingroup$ can you add some references to the 'facts' you present in the question? (i.e. "genomes of prokaryotes tend to have a higher proportion of coding-DNA") $\endgroup$ Commented Feb 7, 2017 at 1:17
  • $\begingroup$ It doesn't take a constant amount of time, but the deletion of a single non-coding sequence is not going to noticeable change that time. more and more functions for introns are discovered everyday, so their may be far less junk than we think. ncbi.nlm.nih.gov/pmc/articles/PMC3325483 $\endgroup$
    – John
    Commented Feb 7, 2017 at 3:11
  • $\begingroup$ I know this is a losing battle, but for the record: Very little eukaryotic non-coding DNA is regulatory. Almost all of it is truly useless, and the simple reason for this is that there are many mechanisms that allow DNA in the genome to expand, and very few mechanisms that can safely remove non-functional DNA; it's therefore more adaptive to ignore the large amounts of harmless but useless DNA than to try to trim down the genome, which would provide very small benefits compared to the risk. $\endgroup$
    – iayork
    Commented Feb 9, 2017 at 13:26

1 Answer 1


A) Here's the thing about evolution - just because something might have an adaptive benefit does not mean that trait will evolve ; this is contingent on the occurrence of the necessary mutations to produce a phenotype in the first place.

B) There are multiple copies of the transcriptional machinery (RNA pol, Transcription Factors) - numbering in the tens of thousands when you are talking of the core transcriptional machinery.


The 3d structure of chromatin , as well as the large amounts of the core transcriptional machinery, prevent there from being too much adaptive benefit to losing large amounts of noncoding DNA.

C) Thirdly, there are suggestions that enhancer promoter contacts, which can greatly facilitate transcription, may be in some part at least determined by sequence motifs in the loops that bring together enhancers and promoters ; so there would be costs associated with losing ncDNA en masse. See www.nature.com/ng/journal/v48/n5/abs/ng.3539.html

D) Indeed, even when one looks at repeat elements derived from endogenous retroviruses, there are those that are of functional significance; Syncitin (which actually codes for a protein) is responsible for trophoblast fusion + placenta formation (See https://www.ncbi.nlm.nih.gov/pubmed/10693809 ) , and there are multiple repeats that themselves serve as enhancers of interferon response genes. See https://www.ncbi.nlm.nih.gov/pubmed/26941318


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .