You can use bioinformatics once you have isolated the gene of your interest and have a sequenced it. Following sequencing, you can perform  BLASTn  using NCBI server, and compare your sequence with those in the database. By this you will even come to know if your sequence is novel or already existing. As mentioned by Qazi, it is difficult to identify novel genes and you should have a precise aim and reason for the study. Nevertheless you can begin with isolating, purifying and sequencing and eventually comparing with the database.

I'm not sure if I understand your question correctly. Are you trying to find ways to discover novel genes or are you trying to identify the functions of novel genes that you've found? From your title it sounds like you're trying to identify function, in which case the two above suggestions are good. DAVID is a really nice tool to use when looking at function and it even produces some graphical output. I would also try a Blastx (NCBI) to search in protein databases using the translation of your DNA/RNA sequence. The reciprocal best hit method is a good way to estimate orthology, and then you can use information about the ortholog(s) to find out more about the potential function/pathway (KEGG/COG analysis etc) of your gene.

follow this article:

http://bib.oxfordjournals.org/content/8/1/6.full.pdf+html

Once you have the sequence you can use functional prediction algorithms such as InterProScan that attempt to identify functional protein domains and assign GO terms.

i have a DNA sequence and also Protein structre and me also predict the 3d protein stuctre but now the tim to find out the novel gene protein function. and my aims is to find the find function of gene which are involved in the fruit repining, qualtiy nd quantity and also find out which genes are protein which work as vaccine.

If have a validate protein structure you find the function by using ProFunc Server: server is to help identify the likely biochemical function of a protein from its three-dimensional structure. It uses a series of methods, including fold matching, residue conservation, surface cleft analysis, and functional 3D templates, to identify both the protein's likely active site and possible homologues in the PDB. You go to ProFunc server through this link. enjoy your aim....

http://www.ebi.ac.uk/thornton-srv/databases/profunc/

Thank u Yogendar Thakur

Hi Tayyba, I guess you are working with a plant genome. You could find orthologues of your genes with the arabidopsis thaliana genes that are annotated very precisely. Once you find Arabidopsis genes that are similar to yours you may infer the function of your novel gene. You could use imparanoid for the orthologous identification and the website Agrigo to get the functions of the found genes.

The gene is novel, right? Knowing the DNA and protein structure is fine, but not sufficient. First and foremost you want to delete the gene and check for phenotypes, before you browse any databases. This can tell you a lot about the role of the gene in modulating cellular functions, but you need a keen eye for detail. You may want to check whether there are macroscopic or microscopic changes in your gene loss-of-function in plants, for instance cell differentiation, apoptosis, cell proliferation. Only of you have a clue on what the mechanistic implications of your gene might be you can start to browse databases in a focussed manner.