I'm reading a text (Wagner, 2007) on identifying positive selection. In the paper, the author says that low complexity regions are known to be associated with the relaxed selection. I'm trying to understand why this is true. I get that if the region is under relaxed selection then most mutations would be tolerated but why would that lead to low complexity (other than something like slippage during DNA replication)

It's in the section: "Little overlap between variation clusters and low complexity regions"

Wagner A. 2007. Rapid Detection of Positive Selection in Genes and Genomes Through Variation Clusters. Genetics 176(4): 2451–2463.

  • 3
    $\begingroup$ can you provide a link to the text that you are reading so that the context of the author is clear $\endgroup$
    May 3 '13 at 4:11

Low complexity regions have repeats in nucleotides ( or amino acids ). E.g. PPCDPPPPPKDKKKKDDGPP or AAATAAAAAAAATAAAAAAT. This 2011 article compares 14 individual Plasmodium falciparum genomes and finds that these repeats are highly variable between individuals - some may be short AAAATAAAA others may be longer AAAATAAAAAAAAATAAAAAAATAAAATAAAA....AATAA even to scores of bases.

I think its not clear what these sequences do, but the mere fact that they can vary so much from one viable individual to another indicates that they are not strongly selective. *If they were subject to selection, you would be able to see the difference in the variation in the health, fitness, and reproduction that these significant genetic changes have, but in this case the change in sequence would not change the properties of the sequence much if at all. *

I think these are also related to copy number variable (CNV) regions. which expand and contract in the genome easily even between identical twins. (sorry for the indirect reference - like many SDN stories, the reference is totally wrong), but I remember this work...


What is complexity?

"Complexity" in this case is derived from the SEG algorithm, which is based on Shannon's information entropy. This is not necessarily repeated regions, but regions in which a limited pool of characters are used (for example transmembrane regions, in which you are far more likely to see a hydrophobic residue than a polar one). These scripts can help you calculate sequence complexity if you have specific examples in mind.

This idea asks how "compressible" is the information in the string of characters (think zip files). Random characters in strings need more information, whereas less variable, more predictable regions can be recorded with less information and are less complex. So a region in which you can more easily guess the next amino acid is simple, and more random amino acid sequences are complex. Sequences with more complexity are often more functionally important and more useful as functional indicators than low complexity sequences. Indeed, BLAST used to mask low complexity regions as they were not useful for homology searches.

Why are complex regions more variable?

The question as I understand it is asking about this fourth conclusion from the paper.

Fourth, variation clusters occur in structurally well-defined and functionally important protein domains of high sequence complexity.

This is now not as paradoxical as it seems. Regions that are highly variable with lots more random/unpredictable amino acids such as those with well-defined structures and functions are more likely to occur in areas with high complexity as a result (which is a measure of randomness in a sequence).

There is more to this than meets the eye...

This is of course not the whole story. In regions of high complexity, there are highly conserved and essential for life residues that cannot be changed. On the other end of the scale, there are also areas like STRs that have low complexity yet are highly variable in length between individuals and underpin DNA fingerprinting.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.