4
$\begingroup$

I am investigating a role of SNPs in DNase hypersensitive sites and in the DNA regions of histone marks and have some questions about it.

SNPs in DNase hypersensitive sites might mean that those SNPs are in enhancer regions and might influence TF binding affinity what can influence a transcriptional output. Is it correct?

SNPs in histone marks sites I have the number of SNPs that fall in the peaks (chip-seq) for the following histone modifications of healthy patients: H3K27ac,H3K4me1,H3K4me3 (represent the active chromosome) H3K27me3, H3K9me3 (represent the repressed chromosome)

Histone modifications represent if the chromosome is active or repressed. A SNP might change the binding affinity of a histone. It means that if I have a SNP in data for active chromosome (H3K27ac,H3K4me1,H3K4me3), it might change the status of a chromosome, namely turn it into a repressed one or it might induce a histone not to bind at all(it means that this region is still active,no?) Is it right? I am not sure I understand the possible influences of a SNP histone marks sites.

$\endgroup$
  • $\begingroup$ Enhancer regions or other protein binding regions of DNA will not be DNAse hypersensitive. $\endgroup$ – WYSIWYG Apr 1 '15 at 9:06
  • $\begingroup$ @WYSIWYG Why? I though it chromatin is not repressed, then it is nucleosome free. $\endgroup$ – Alina Apr 1 '15 at 11:51
  • $\begingroup$ Protein bound regions will be protected- same principle as in case of nucleosomes. $\endgroup$ – WYSIWYG Apr 1 '15 at 12:25
  • $\begingroup$ Why will it be protected? In case the TF did not bind to it, it will not be protected. I have seen lots pf plots where DNA hypersensitivity coincides with TF binding sites (chip-seq peaks). What is then hypersensitivity sites and why do we want to explore it? Wiki suggests: "These accessible chromatin zones are functionally related to transcriptional activity, since this remodeled state is necessary for the binding of proteins such as transcription factors." $\endgroup$ – Alina Apr 1 '15 at 12:39
2
$\begingroup$

First off let's define some concepts.

  • DNAse hypersensitive regions are DNA regions which are in an open chromatin conformation (i.e. euchromatin). This means that those regions are more active at the genomic level (i.e. higher gene expression, gene regulation and higher TF binding) and are less prone to form nucleosomes.

  • Histone mark sites are DNA regions known to bind certain type of histones and therefore also influences the chromatin conformation leading to the same effects described previously. The critical point is that the histones you are listing do not primarily bind a specific DNA sequence (as described by @mdperry) but rather DNA sequence marked by epigenetic modifications such as methylation and acetylation.

Now for your first questions, SNPs in DNAse hypersensitive regions might influence the chromatin state and therefore yes might influence gene expression and gene regulation. Will those SNPs also influence TF binding? Well they might indirectly via chromatin modifications. They will not directly act on TF binding if not located in a sequence known to bind TFs (TFBS). DNAse hypersensitivity is not a synonym for transcription factor binding regions/sequences.

For your second question, SNPs in histone mark sites are a little bit more difficult to interpret. They might, or might not, provoke chromatin changes by influencing the binding of histones but not in a direct way. CpG islands are, for example, DNA regions known to be greatly under the influence of epigenetic marks and therefore mutations in those specific regions (e.g. a C -> T mutation) might influence the relative methylation/acetylation state and therefore the binding of histones. SNPs might therefore influence gene expression via chromatin changes which leads to higher/lower TF binding but again not in a direct way. Here again a SNP would influence TF binding directly only if the mutation is located in a TFBS.

As you can see my answer is full of "might" and the reason is that it is very difficult, at the current state-of-the-art, to determine the exact impact of a SNP on the 3D conformation of the chromatin and therefore predict the effect on specific gene expressions.

An advice would be to try to co-localize your SNPs with known TFBS either predicted In-silico or measured via CHiP-SEQ experiments and merge that with the information about DNAse hyperactivity and histone marks you already gathered. For example a SNP in a TFBS in a DNAse hyperactive region is very likely to have a greater impact than a SNP in a TFBS in a non active region.

I am sorry I cannot be more specific but this is a very active research area in genetics and there is still a lot to understand, especially for the interactions between all the players influencing the chromatin conformation.

$\endgroup$
  • 2
    $\begingroup$ (at)Tonja, you could search far and wide on the Internet and you would not find a more cogent or on point discussion of these concepts than the one @cagliari2005 very patiently elucidated for you. $\endgroup$ – mdperry Apr 1 '15 at 18:32
1
$\begingroup$

Keep in mind that it is not individual histones that are binding, there are 8 histone proteins forming a compact nucleoside particle, and the particle has 146 bp of DNA tightly wrapped around it. Since the interaction between the histones and the DNA are not sequence-specific, it is unlikely that a single nucleotide polymorphism would have a measurable affect on the occupancy of the nucleosome at that site.

$\endgroup$
  • $\begingroup$ "interaction between the histones and the DNA are not sequence-specific". So, in fact the more SNPs are situated in 146 bp region, the more effect they have? I have lots of papers which refer to SNP in histone sites but non of them tells about the consequences. $\endgroup$ – Alina Mar 31 '15 at 20:57
  • $\begingroup$ No, you are approaching it like a classical lac Repressor <-> lac operator interaction where you can measure the moles of free protein, free DNA, and bound protein-DNA complex, and use those values to calculate the Kd, and a binding affinity. That model is for sequence-specific DNA binding proteins, and yes, the more nucleotide sequence mismatches you introduce into the recognition site, the weaker the affinity. The histone octamer core will bind to any DNA, and is by definition not sequence-specific. So introducing SNPs is unlikely to affect occupancy $\endgroup$ – mdperry Apr 1 '15 at 1:10

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.