Hello Mr. Azeem,

Molecular fingerprints of diversity can be achieved by coupling PCR amplification of taxonomical targets with sequence dissimilarities analyzed by denaturing gradient gel electrophoresis (DGGE) enabling the separation of DNA fragments of identical length but of different sequence, resulting in a DGGE fingerprint that is specific for the sample site. Although no standard tool exists, a variety of comparison tools are available to compare fingerprints and include visual inspection, cluster analysis, nonmetric multidimensional scaling and moving window analysis, which allow several quantitative and qualitative biodiversity indices to be recovered. The Shannon diversity index is a quantitative index which computes band intensity with species abundance, while the estimation of species richness provides qualitative data relating the total number of bands observed to the total species number and species richness and corresponds to the total number of unique biomarker sequences within the ecosystem. PCR-DGGE is capable of detecting as much as 95–99% of the bacterial community, and several candidate sequences can be used as genetic biomarkers including variable regions within the 16S ribosomal DNA gene, the ribosomal DNA gene, and the rpoB gene (encoding the β subunit of RNA polymerase), which are universally present with highly conserved flanking regions allowing the simultaneous amplification with ‘‘universal primers’’. Due to the length and sequence of polymorphisms, the choice of primers, their specificity and degree of mismatch, the reproducibility of extracting DNA from environmental samples with complex and diverse microbial communities and natural materials will affect the recovery of biodiversity indices. PCR bias resulting from inhibition by coextracted impurities or preferential amplification of certain species by various primer pairs is also a key issue. Other important issues include the degree of genetic polymorphism within the target regions within the community, comigration of PCR fragments from different species in the same DGGE band, and the formation of multiple bands during amplification of genes from a single genome. Application of DGGE methodology also has size limitations of 100–500 bp, which may affect the probe design and further phylogenetic analyses, while subsequent quantitative analysis of the DGGE fingerprint obtained assumes that the number of bands reflects the microbial diversity in the sample and that their relative intensity reflects abundance. 

Here are some of the links which will help full to you..

http://download.bioon.com.cn/view/upload/201111/29125905_9842.pdf

http://books.google.co.in/books?id=k_-0-DWxuRgC&pg=PA245&lpg=PA245&dq=protocol+for+good+bands+on+DGGE+gel+using+soil+DNA&source=bl&ots=YogMqZ8Tk2&sig=D-Bbz45ZN8RYCefML_8CQo45Ytc&hl=en&sa=X&ei=dwF3VK2BB9W1uQT_k4KADA&ved=0CDkQ6AEwBQ#v=onepage&q=protocol%20for%20good%20bands%20on%20DGGE%20gel%20using%20soil%20DNA&f=false

http://www.agriculturejournals.cz/publicFiles/11516.pdf

http://ddgehelp.blogspot.in/2005/11/dgge-guide.html

It would be a good idea to quantify each product and normalise the amount of DNA you load into each well. Hypothetically, if you had a strong PCR product that only split into a few bands and a weak PCR product that split into many bands you'd get very uneven signal strength. If you were using a GelDoc to visualise the gel, you might find that the auto-exposure hides your weaker bands.

Also, what stain were you using? Techniques like DGGE and SSCP give much weaker signal than a normal PCR/agarose gel so need a more sensitive stain like SYBR Gold or silver nitrate. This is related to my first point, in that some strong bands might be visible under a weaker stain like SYBR Safe, but other bands might be impossible to detect.

Finally, do you know anything about the community composition of your samples that aren't working? Is it possible that your gradient isn't strong enough to denature these samples and they're just running off the gel?

You can purify your soil DNA and do the pect with your primers of choice but with GC clamp and then seperate it on DGGE. You will good band profile.

Thanks David, i am studying 16S. Yes it looks like problem is also in gel doc. But some bands are more dark. David can you tell me what will be the concentration of silver nitrate for staining? Is it necessary to load equal concentration of sample DNA? The DNA concentration which i load were in the range of 13-22 ng/uL(20 samples)

You might try experimenting with your denaturing reagent gradient concentration. Start weak and keep increasing your concentration on different gels. Try a 20% top 50% bottom gel to visualize a different soil microbial community, then, keeping the ratio the same, keep increasing both the top and bottom concentrations. Sometimes the ratio increments can be too wide, not allowing visualization of parts of the soil community. By using the same marker of known organisms in each incremental gel, you can then piece the images together.   As stated by Sunil S Adav, a GC clamp enhances your banding profile. David's suggestion too, is of note, the denaturing reagent concentration may need to just be increased.  If you are achieving good PCR amplification, then I would suggest the gradient concentration experiments. 

If you're only doing qualitative analysis then the concentrations don't matter too much (although obviously the lower they are the more risk of missing bands you become). For quantitative analysis, where you interpret the band intensity, then I think the concentrations are much more important. I don't think this is likely to be your problem though, since you've got good PCR products. Silver nitrate stains usually use a 0.1% solution, I've attached a few papers on protocols.

Kelly probably has the best advice though, start with a very low denaturing threshold just to see if you can see any bands. Once you know you've got product you can up the gradient to get better resolution.

http://www.coulterflow.com/bciflow/singlecell/protocols/EP10002_PAGE_process_instruction_s.pdf

Article Silver Staining of Denaturing Gradient Gel Electrophoresis Gels

Article A comparison of silver staining protocols for detecting DNA ...

I would suggest you to load different concentration of pcr of product (lower to higher) of only one sample (in which your getting good amplification) on dgge gel and follow the silver staining method to visualize the gel bands as it is very sensitive method ranges from pg to ng. once you standardize the concentration of pcr product required to load on gel then use the same concentration of pcr product  of all samples to load on gel. Before proceeding towards dgge gel always confirm that sample is getting amplified correctly followed by gel quantification of pcr product and then accordingly load the volume each pcr product on dgge gel. Some times bands will not be separated properly on dgge gel. I n that case try different  gradient of denaturing % and as well as page gel % (acrylamide : bis acrylamide). good luck....