If the bands are running very close together, it's likely that the protein becomes post-translationally modified: phosphorylation, glycosylation, etc, may occur and may slightly change the mass of your protein.

Hi Vimarsh, Will you clarify a few questions that I have? Was this observed in a reducing or non-reducing gel? what proteins are you trying to check? does it have two different sub-units and if so are they supposed to form a hetero dimer?

in some cases the sub-units form a homo dimer resulting in a band almost closer to the expected one. Subject the sample to a size exclusion chromatography for further confirmation. If the band is running higher than the expected size after purification then there are aggregates of the protein you have isolated.

What kind of staining do you use? I have seen this with the Siverquest staining kit where overstaining can have the effect that the signal extighuishes in the middle of the band thus resulting in 2 bands. In the middle it appears yellowish than. Loading less hepls. Hope this helps  

@ Paul, We use Coomassie Blue staining. As markers are clean, i dont feel its a problem with staining.

@ Mallika Natraj,  As being a reseach protein, we dont have much detail on it. However, Bands are very close and lower this thicker and upper is bit thin in nature, earlier PAGEs for same protein have crisp bands, but second bacth is showing duplex pattern.

@ Daan, You answer seems to be matching to my conditions hereon. But i am surprised to imagine such a changes in band size due to PTM. 

Some minor proteolysis can occur even in the presence of SDS, try lowering the SDS to near its limit (0,1%) incrementally, until the bands disappear.

Have you done a western blot? That will verify your protein band.

@Mallika we have not done that, we will work on it. As of now, the question remains a mystery to us.. :)