You could  gel-purify the PCR product and send it for sequencing together with one of the amplifying primers. If the sequence comes clean, it means that you only have one species; otherwise, you will get overlapping traces for the different nucleotides in the chromatograms

Try pre screening your primers on a gel including a temp gradient from Tm-5C to Tm + 1C in 1C intervals

Perform this for 2 x sets of primers and take the primer set plus actual specific and efficient Tm into the qPCR cycling reaction itself

Having done this add a melt curve cycle to your customised qPCR reaction: A single peak or two converging peaks indicate a single product (converging peaks can result from secondary structure in your sample but are non the less derived from the same amplicon)

As controls, include an NTC control; that is water instead of cDNA: If you obtain a peak in this sample that indicates primer dimers or even primer oligomers (as primers are the only source of DNA)

Also include a no RT control: That is an RT reaction where cDNA has been replaced by water. Include primers in this reaction. Peaks from your no RT control indicate contaminating genomic DNA

Going forward and to make things more specific, apart from customising your annealing temp to render qPCR specific and also including a melt curve cycle to verify extra peaks, add 5% DMSO to your qPCR reaction. This will minimise formation of primer dimers. In addition, reduce your primer concentration if possible from 5-10uM (if using) to 2.5uM or even as low as 1uM: Even if that increases your Ct value, providing that value is < 30cycles and triplicate technical replicates are within 0.5 cycles your data is sill valid

Find attached a stratagene guide which in 10 years of performing qPCR I still think is the best   

You can also use a high concentration agarose gel and run out your samples under low voltage to see if you get multiple bands.

To use just a melt-curve analysis, add a fluorescent dye to your reaction (SYBR green, EVAgreen, etc).  You can buy just the dye and it is much cheaper than the PCR mastermix with dye and enzyme, etc.  Use the melt curve settings and make sure you have a FULL set of controls (no template, positive control, negative control).  Note: it is possible to have a single PCR product that gives a double peak.  Make sure your products are small, less than 200 bp if possible.

You could  gel-purify the PCR product and send it for sequencing together with one of the amplifying primers. If the sequence comes clean, it means that you only have one species; otherwise, you will get overlapping traces for the different nucleotides in the chromatograms

Use hot-start or nested PCR.

If you have a restricted tian site in your target the product should digest to predicted sizes.

Hi Dong, I have some suggestions or comments in this, and do hope would be helpful for you.

Before switching towards the analysis or any conclusion about the presence of non-target products you need to be sure that you have followed the following strategies, which should be done or looked earlier than qPCR and some while performing your qPCR, so you can have a definite assurance that the products you found are really nonspecific unique ones. These include;

You must perform primer efficiency and select the best concentration of them, which gives maximally distinct Tm.

You must perform DNAse treatment to rule out (at some extent only) the possibilities of gDNA so not to get nonspecific peaks to get confuse later on.

Observe the NEC control wells status (should be with negative amplifications). [Wells with your extracted RNA]

Primers must be designed to skip introns (span exon-exon boundaries).

Never design large products, keep them low ideally within 70- 100 bp (to avoid secondary structures or any chances of detection away from your region of interest from your primers).

You must keep in mind that a peak height along with its sequence and length is also highly dependent on the amount of product in your reaction well (a fact often ignored or not taken seriously), and it is due to improper pipetting between the same triplicate set, use of non-calibrated pipettes, poor sealing of plates so chance of sample evaporation is always there, delay in loading the plate into machine. These all result in various sized peaks under the same Tm, which are of same product, but amount differs in each reaction.

Majority of the issues related with multiple nonspecific peaks can be resolved by taking the above facts into an account. And you should know that primer dimer peaks can easily be resolved even by a look on melt curve and also the gel picture will give you this idea, as they are very small usually near 50 bp or so, dimmer and in last at your gel.

Also after the qPCR run, if you just find multiple bands on gel, along with multiple products’ peaks on melt curve then it has some logical explanations like;

If different peaks under a same Tm, so may be the Tm is so close enough that melt curve cannot resolve or analyze them.

The programmed or default range of melt curve in your machine does not include all Tms for all the products appearing in your assay. For example, if the melting starts from 70 C so by the time, except any one, all others products already become single stranded, thus on gel appear as near-sized separate bands.

If you see shoulder peak linked with your product peak on melt curve, it just reflects some average complexity region in your product, which results in this kind of non-uniform melting. So don’t consider it as some non-target product.

Heterozygote polymorphism in your GOI also results in 2 distinct peaks of a same product. So HRM (high resolution melting is one choice to distinguish if you have the facility and resources in your lab, as the HRM Tm shifts are even less than 0.2 C. So it provides an excellent resolution).

Finally, in my opinion, a more easy and handy approach is to add more steps in your melt curve program, as various machine offer this, you can here select Tm ranges as well for each added step. So it’s a useful approach to detect various non-target products having different Tms. Just have a look on the melt curve program in your machine.

With good wishes…