I would advice you not to use a single locus or either multimarker analysis for species definition. It is highly disputed item. Species is often defined as the group of organisms capable of interbreeding and producing fertile offspring, so, it cannot be defined solely based on molecular analisys, and there are so many plants with different flowering time and morphology still from the same species...

That is Biological Species Concept and i am favor of it. As the definition goes Group of similar ind of individual that freely interbreed and produce fertile offspring and are Re productively isolated from other such similar groups. B ut the molecular taxonomist not think this way to them there is nothing like species , genus , class or order, or kingdom concept for them only clades and sister c lade and taxon. Its unacceptable. But now The LINEAGE species concept is on way to tackle the thing to describe a species to be distinct this concept say that integration of morphological,characters, molecular similarity and even the niche modelling is necessary other summative informative is needed to describe a species. this taxonomy science is going to be tedious for sure.

So using multiple markers can definitely give better results but for more precise results I would suggest you can use multiordinate approach by using molecular, morphological and ecological data (niche modeling). This way you will have more confidence in your results and also most journals now accept papers with multi disciplinary approach.

And you have just blasted one gene; it will just show the similarity across species for that particular gene. As mentioned in above answer any gene will show you similarity with many different species. I would suggest you do a concatenated analysis (i.e. combine all your sequences from different genes) and analyse it with as many different species as you can find on NCBI. Then you will have some idea regarding how different that species is from others.

I totally agree with you kunal

ok, that means initially its was showing 99% homology in blstn means that may or may not be a new species. So, how do I know from sequences that it may or may not be a new sp? What are the procedures by which I can understand up to species level, starting from handling with raw sequence data.

Kindly give me your valuable advice regarding raw sequence editing and sequence alignment as I want to know this first. Then step by step single marker analysis and multimarker analysis.

I'm assuming that you want to use an evolutionary species concept, which is the one that is really assumed when we do population or evolutionary genetics. It includes the BSC in the case of sexual organisms. If you want to find evolutionary species, you can't use just the sequence similarity of specimens. Instead, you have to compare sequence differences within each putative species to the sequence difference between them, OR use various tests for gene exchange. The former can be done with the K/theta ratio; see Birky 2013 PLOS One 8(1):e52544. There are a number of methods for the following, many of which are referenced in my paper. You might also be able to use the GMYC method that is also referenced in my paper, developed by Tim Barraclough and his group.

@ Birky: you meant this article: "Species Detection and Identification in Sexual OrganismsUsing Population Genetic Theory and DNA Sequences (2013)"?

As i learned, when BLAST query sequence, below 86% similarity are different species. Anyone agree with this?

Yes, that's the paper.

Thank you Sir for your valuable information. Yes Sir, I really want to discriminate between species. I have collected different morphotypes across its global distribution. Among them last year one individual I have collected after 30-35 years reported in one paper. But in all other papers, it remained as synonym. The 1st one occurs in different areas of the country(3000-5500 masl) but 2nd one occurs in a area where the 1st one and 2nd one occurs together. Flowering time is also different. I have got one more individual from alpine medew having completely different flower color.

So, what would be my approach to discriminate those species??

Surely multimarker analysis might help to see the difference, provided you can analyse the homology data from any local alignment tools like BLAST.

You will want to build a phylogeny from your sequences before you take 99% Blast identity (technically, homology is a "yes" or "no" term and cannot be a percent) to represent belonging to (or not belonging to) a given taxon. Blast only assignments have caused problems in the past, as this article highlights (while it is focused on bacteria the points should still be applicable to plants): http://www.ncbi.nlm.nih.gov/pubmed/11443357

Two closely related plant species can have very similar chloroplast gene sequences. If working with closely related taxa you should be using intergenic regions from the chloroplast to provide more phylogenetic information. Remember that chloroplasts are generally maternally inherited, if the two species are a diploid and polyploid, they could share the same chloroplast haplotype. ITS has many inherent problems but can give some initial clues for phylogenetic analysis. What genus or plant family are you working in? Check NCBI for potential "single copy" nuclear genes to round out your multi-marker analysis.

Thank you Mam for your valuable suggestions. Mam, I am working on some Indian representatives from family Valerianaceae order Dispacales. I have amplified rbcla, matK, trnL-F and psbA-trnH intragenic spacer already. I am now trying to amplify ITS1 and 2. I did not analysed the results yet. Kindly give me some suggestions if I have to amplify any other region, Mam.