While it will be very difficult to assign importance to SNPs within intergenic regions unless they reside in previously identified enhancers, some weight can be assigned to intronic SNPs. It's been previously shown that intronic SNPs within 200 base pairs of the splice donor or acceptor sites can have functional effects on splicing. This is particularly true if the SNP occurs in sequences that are normally GGG. Obviously, changes to the splice donor (GU) or acceptor (AG) sites will cause the biggest changes to splicing. Additionally, look for changes in polyA or polyG (or polyAG) sequences in the introns as those are important for lariat formation. Several papers have been published with specific intronic sequences that are important for enhancing or suppressing splicing. I would start by looking to see if your SNPs fall in regions with similarity to any of those sequences.

I´m not an expert, but maybe if you pick the ones localized next to the intron/exon junction, the chance to affect splicing is higher.

Leonardo makes a good point.

In addition, you may look for regulatory elements such as miRNAs that may reside in introns and intergenic regions.

In intergenic regions there may be many thus far unknown regulatory elements (see ENCODE). So there is really no easy way to prioritize the intergenic candidates. Maybe you could align with human regulatory elements and analyze better matches first.

While it will be very difficult to assign importance to SNPs within intergenic regions unless they reside in previously identified enhancers, some weight can be assigned to intronic SNPs. It's been previously shown that intronic SNPs within 200 base pairs of the splice donor or acceptor sites can have functional effects on splicing. This is particularly true if the SNP occurs in sequences that are normally GGG. Obviously, changes to the splice donor (GU) or acceptor (AG) sites will cause the biggest changes to splicing. Additionally, look for changes in polyA or polyG (or polyAG) sequences in the introns as those are important for lariat formation. Several papers have been published with specific intronic sequences that are important for enhancing or suppressing splicing. I would start by looking to see if your SNPs fall in regions with similarity to any of those sequences.

While not directly answering your question, linked variation in intronic regions may have important deleterious effects in some systems: http://rspb.royalsocietypublishing.org/content/276/1657/657.full

Could look for known autolytic motifs first. The nascent RNA strand will need to fold back and anneal with another region of itself to form ribozyme, you can use the same idea of melting points for primers to assess the annealing capabilities of different regions forming base pairs and localized charge estimates to gauge ability to recruit inorganic ions as co-enzymes for unpaired bases in loop domains, A R Krainers "Eukaryotic mRNA processing" may be a useful read.

SNPs in exon/intron bundries are the most important region to be regulatory. As friends said you could align your sequences to other species or use prediction tools to find SNPs affecting regulatory regions. Mutationtaster could help you in this way.

Dear Samra

Its an ever-interesting and a debatable question which specifically have no answer as such but we have a lot of speculations about which need experimental evidences. The answers so far we have are based on certain evidences and have strong indications that the non-coding regions have regulatory functions for the coding regions, means the coding regions are the lead heroes on the stage but the direction is being done back stage and don't have direct visibility for thier work.

Yet they are the heart of the system. Thus any modification in non-coding would definitely have impact on the coding expression and regulation. for instance:

SNPs specifically at splicing region may prevent splicing, SNPs in intergenic/ promoter/ enhancer/ silencers may lead to improper and defective expression and function,and so on. but i think the splicing one would affect the structure as well as function of the protewin whereas the SNPs in other regions would affect only function of the protein but have lesser or no effect on protein structure.

If you have the genome information, you can amplify your region of interest (splicing regions) using different sets of primers, clone and sequence them (sangers) and then align against the database.

Yup in order to narrow down, you can use prediction softwares.

Thanks a lot for all positive contributions. Aleksandra, I am working with mouse model (CIA). Your point about splicing rate is very interesting, at least for me as I did not know about that. Can you explain a bit or provide a reference?

Dear Marisol,

I have a long list of SNPs from NGS so I don't think it will be practical to carry out mutagenesis and protein expression studies for all the SNPs. I would like to prioritize them before I do such experiments.

Daer Samra,

It is not easy to consider a specific priority to select a/some particular SNP(s). One of the best was is to consider intronic regions which are involve in splicing process, including Acceptor an Donor sites of splicing sites, Consensus region nearby splicing sites an finally the branch site in the middle of the intron.Moreover you could find the exact sequence of theses regions/sites in genetics references like Strachan.

Good

luck

Hi Samra,

You have a lot of great advice above about splice sites but sometimes conserved and important elements are not so obvious, especially in intergenic regions. Two ways to try to identify these elements are both bioinformatic and therefore usually high throughput. First align regions that have SNPs with the sequence from very closely related species. Identify conserved regions that contain SNPs. By adding increasingly less related species, you should be able to narrow down the regions of the greatest conservation (above background). In a region with strong conservation across diverse species, evolutionarily speaking, there should be something that is functional and being conserved by natural selection. The second way you can look is to use a database of conserved regulatory sequences and search your region for these motifs and see if any of your SNPs fall into these know regulatory sequences and binding motifs. Combining the search for evolutionary conservation with identification of known regulatory sequences or binding motifs can be very powerful. Any findings can usually be confirmed in tissue culture cells. Hope that helps!

Hi,

In addition to the previous excellent comments, maybe identified SNP haplotypes (if any) may help you select only some of them.

can you look for phenotype characteristics in your region of interest, i was looking for some articles and found a study on dog breeds http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002316 may be you find some interesting ideas

If the strains you sequenced are among the more commonly used inbreds, you can use the SNP resources at the Mouse Phenome Project (http://phenome.jax.org/) or Sanger's Mouse Genomes Project (http://www.sanger.ac.uk/sanger/Mouse_SnpViewer/rel-1303) to validate whether your SNPs have been previously observed. It won't help with functional annotation, but may save you some time and effort if you are just looking to identify true SNPs and potential NGS errors.

If you have SNPs within genes and haven't yet identified your most promising gene candidates, you can use the Mouse-Human Disease Connection from MGI (http://www.informatics.jax.org/humanDisease.shtml) to pull out all of the annotated phenotypes to all of the genes in the region (search by genome location). If any of them look relevant to inflammation or autoimmunity, they should probably go up to the top of your list.

Echoing others replies, a including a mixture of criteria such as spice acceptor/donor sites, sequence conservation, as well as experimental evidence such as TFBS would be worth investigating. A good place to start is the UCSC Genome Browser for the mouse genome: http://genome.ucsc.edu/cgi-bin/hgGateway?hgsid=367936567. Depending on the number of SNPs you plan to test, you can restrict or relax criteria to exclude or include more variants, respectively.

Here is a recent paper using a similar approach like the one described by Heather to identify SNPs/variants in non-coding regions affecting TF binding and hence the expression level of associated genes: http://www.ncbi.nlm.nih.gov/pubmed/24439387

Hi Samra,

To prioritize the SNPs, we may first to look the genetic model RA mouse (either is correct or hypothetical). It is much like to study the single locus disease or multifactorial disease (like obesity). For the former, you can concentrate the nonsense SNPs or causing alternative transcripts, however for the later, all the SNP can be important even outside the gene annotation regions, as ofter long distance enhancers which regulate gene function are not well annotated.

At the protein and post-translational levels, variation in protein stability due to SNPs in coding sequences can cause different levels of enzyme activities.

SNP in a regulatory DNA binding site may alter the affinity of the regulatory protein such as transcription factor (TF) resulting into different mRNA levels. The SNP in promoter region leads to decreased affinity as shown above, resulting in low gene expression. Several examples of the effects of SNPs on mRNA, translational levels and post-transcriptional levels.

Full text and reference: http://www.sciencedirect.com/science/article/pii/S1110863012000687

One more thing to add. Besides prioritizing the intronic or intergenic SNPs one by one, I would also examine the local linkage disequilibrium (LD) structure to see how these SNPs relate to each other, and to the other nearby SNPs located in exons/promoters/miRNA binding regions. An intronic or intergenic SNP may look very significant for your model but that could be due to its location in a LD block in which some other SNPs might be the underlying factor.

1) Intronic sequnces contains branch sites necessary for properly splicing and mutations could activate criptic splicing sites[1] ;

2) The taxonomy intron/exon derives from an annotation (that could be wrong). You should run RNA-seq sequencing to assemble the transcriptome and align it to the genome, in that way you will be confident of the splicing patterns

[1]http://en.wikipedia.org/wiki/RNA_splicing#spliceosomal