It would be unpractical or difficult to experimentally address the function of every mutation seen in non-coding regions.  I would assume the first line of approach may be a computational one. If a SNP (or other mutation) is reported as not greatly damaging (i.e. it is listed in the 1000 genomes project dataset) then there would seem little justification for setting up an experiment to test that mutation. If the mutation is not reported in the 1000 GP, then other mutation datasets (such as HGMD  - if you have access to it - or others) may be checked. If it is not listed there either, then a third approach may be to check if the mutation falls within a genomic region with active chromatin marks (such as those listed in the UCSC human genome browser), in which case it might disrupt an enhancer (a case that may be followed-up experimentally perhaps by cloning that region with a gene reporter system). Other cases in which a mutation may interfere with critical secondary structures of 3' or 5' UTRs may require chemical probing of RNA molecules corresponding to these (or part of) UTRs with and without the mutation, which may give some hint as to their folding (and ligand binding in the case of riboswitches), at least in vitro. 

mutations occurring in untranslated regions may enhance or suppress a binding site for transcription factors (this happens for several milk proteins, which share the same primary sequence but depending on the allele may have different expression levels). Another example is given by selenoproteins which require the presence of one (or more) selenocysteine insertion element in the untranslated region for the insertion of the aminoacid Selenocysteine. If this element is not present, the process can only result in a truncated inactive protein.

Dear Leonardo,

Thank you very much for your answer, I understand that it has a regulatory effect, but how to define that effect and how to confirm it? which technique should be used? These  are the important questions that still searching for answers!!!!

It would be unpractical or difficult to experimentally address the function of every mutation seen in non-coding regions.  I would assume the first line of approach may be a computational one. If a SNP (or other mutation) is reported as not greatly damaging (i.e. it is listed in the 1000 genomes project dataset) then there would seem little justification for setting up an experiment to test that mutation. If the mutation is not reported in the 1000 GP, then other mutation datasets (such as HGMD  - if you have access to it - or others) may be checked. If it is not listed there either, then a third approach may be to check if the mutation falls within a genomic region with active chromatin marks (such as those listed in the UCSC human genome browser), in which case it might disrupt an enhancer (a case that may be followed-up experimentally perhaps by cloning that region with a gene reporter system). Other cases in which a mutation may interfere with critical secondary structures of 3' or 5' UTRs may require chemical probing of RNA molecules corresponding to these (or part of) UTRs with and without the mutation, which may give some hint as to their folding (and ligand binding in the case of riboswitches), at least in vitro. 

You should consider the possibility that the mutation is neutral in effect and is the result of a random process. For example, if you treat cells with a mutagen, you will probably see mutations in numerous locations, some of which may have no effect.

Researchers were not able to measure the role of non-coding regions of the genome until now..............

Mutation in noncoding region

 Types  and function of non coding DNA see wikipedia

http://en.wikipedia.org/wiki/Noncoding_DNA

Mutation in noncoding region vary from SNP (Silent mutation) to chromosomal mutation ......according to the site of mutation... the function affected......

The significance of noncoding DNA mutations in cancer was recently explored......

http://www.slideshare.net/AustralianBioinformatics/investigating-the-role-of-non-coding-mutations-in-cancer-dilmi-perera-27982803

http://www.sciencedaily.com/releases/2014/07/140723131403.ht

Sequencing and other method for mutation detection could help for mutation detection

SNP and mutation detection

http://www.dnabaser.com/articles/SNP/SNP-single-nucleotide-polymorphism.html

Non coding RNA and mutation detection using comparative sequence analysis

http://link.springer.com/article/10.1186%2F1471-2105-2-8#page-1

Genome-wide analysis of noncoding regulatory mutations in cancer

http://www.nature.com/ng/journal/v46/n11/full/ng.3101.html

Mutation rate Coding/non coding

http://www.quora.com/Do-mutation-rates-for-coding-regions-of-the-genome-differ-

from-mutation-rates-for-non-coding-regions

http://seqanswers.com/forums/showthread.php?t=28647

Thank you all for your replies.

Further clarification for the problem I am facing. I am a studying a gene that is known to be associated with a human congenital disorders, after sequencing all the exons of that gene in the affected individuals, I found 2 new mutations in the only non-coding exon of that gene in all the affected individuals and it was persistent  after repeating and optimizing the screening conditions; for me it is not enough to associate them to the disorders, I endeavor to define their effect; Can any one advise me on the proper way to do that ????

You face one of the most acute problem of clinical genetic but before going further, there some issues that need to be clarified.

let state that there is no technical issue and that these two "mutations" are true

you stated that you have two new mutations ? Are sure that these are new mutations ? did you check if it is not a rare SNP in the human population ?

what is the status of these variation , homozygotes or heterozygotes

are those two"variations" identical ?

are the two patients related ?

what the status of the other members of the family for these "variations"

as long as you will not have these info, you will not be able to go further 

If mutations in the gene in question lead to a measurable phenotype in cultured cells, or there is an animal model for the disease, you could explore the function of these new mutations by introducing them into wild-type cells or mice.

Dear All thank you for your supporting replies.

Dear Thierry,

Please, clarify what do you mean with "rare SNP in the human population"

We have two identical variations in 16 patients.

They are non-related affected individuals  (2 brothers are included, we did not study whole families, yet), but please, notice that there is a very high degree of consanguineous marriages in Saudi Arabia  and the individuals are from different clans.

So,  how do you purpose to go further?

Dear Samer,

can you give us a brief outline of the pipeline/protocol you used in order to identify these mutations? did you filter-out known (common) SNPs? did you sequence healthy individuals from the same families? is this disorder dominant/recessive? autosomal/sex-linked? is the mutation homo/heterozygous?

Hi Samer,

I would check out the following websites below.  It is for the ENCODE project at UCSC.  The information from this project can be displayed in the USCS Genome Browser and more can be accessed directly from the ENCODE website .  I would check to see if your two mutation located in the UTR are actively involved in transcription regulation.  If not then they could also be important in the mature mRNA for the regulation of translation.  For your gene of interest has the relative gene expression been tested in the affected individuals with the mutation compared to a unaffected control lacking any mutations in the gene?  A similar experiment can also be done at the protein level as well.  Hope you find this information helpful.             

http://genome.ucsc.edu/ENCODE/

http://genome.ucsc.edu/

Dear Rocco,

I used the sequencing of the 6 exons of the gene that is reported to induce that disease if mutated. all the reported SNPSs are on the coding exons. I did not sequence healthy individuals from the same families, yet. The disorder is recessive. The gene is on X and Y chromosomes. the mutation is homozygous .

Dear David,

Thank you very much for your answer, unfortunately, it is a homeobox developmental gene which is active at the embryo stage.