This is an active area of research and hasn't been well answered yet. We'd assume that genes and gene products with similar functions generally have similar patterns of interaction in PPI networks and signalling pathways, but sometimes neofunctionalization may complicate things. Even if we'd expect them to produce the same or similar PPI, the post-duplication genes and proteins may be expressed at different levels, causing them to produce different results in PPI screens. 

It depends on divergence time, evolutionary constraints on duplicates and cellular context.

could I get some interesting references related to this... I m working on a set of paralogous genes and it would be great to know some interesting tools to analyse paralogs with very less polymorphisms... 

Thanks

Hi Ashmit,

first of all, it depends on how much complete was the gene duplicated and whether or not all regulatory sequnces up- or downstream of the gene were duplicated, too.

For larger duplications, e.g. in segmental duplications where several gene loci were copied, the very same function should be assumed. From then on, evolution will affect both allels on both gene loci, so you have a copy of four identical genes in your diploid genome. It is accepted that this will decrease selection presure on all copies, which in turn should increase the observed mutation rate. As soon as either expression domains or gene function diverged sufficiently to be stabilized through positive selection, two paralogs will be maintained.

In those cases where expression domains evolved faster than the gene's coding sequence, you can assume that identical  protein-protein interactions or protein functions will be present - as long as the cell has the same physiological competence than prior to the duplication, e.g. your interactors are still present in the same cell and the cells have the same plasticity to respond to your paralog.

Besides the hallmark principal articles on orthologs and paralogs by Koonin, there are very few articles that address this issue. Most of them are related to paralogs in  large segmental duplications, where transcriptome data were analyzed.

Moreover, articles are very specific for the different organism, e.g. there are few  articles addressing this issue in the model plant species Arabidopsis thaliana.