My preference for IC50 measurements is to do a series of 2-fold serial dilutions, covering a 500-fold range of concentrations (12 wells, including a zero compound MAX and 100% inhibition control MIN, in a row). Other people prefer half-log dilutions to cover a wider range with the same number of dilutions. I wouldn't recommend spacing the concentrations any farther apart than that if you want to calculate IC50 and Hill slope.

If your supply of materials is severely limited, then pre-screening the compounds at a single concentration makes sense. The concentration you choose will depend on how potent you expect the compounds to be. If you choose too high a test concentration and get 100% inhibition for all 10 compounds, you haven't accomplished the goal.

If materials are not so limited, then I would recommend doing the IC50 measurements right from the start. You will need more than one plate, of course. If some of the compounds have IC50s below the lowest test concentration, you will have to test them again at a lower range of concentrations if you want to measure the IC50.

If you need to save materials and have a suitable plate reader, you might be able to switch to a 384-well format and reduce the assay volume.

So each compound takes 10 wells with max and min being same. Do you take single double or triplicate reading? Is single reading seen as bad or low quality data?

I am starting to see ur point. with 96 well i cant have IC50 for many inhibitors in tripllicate readings. What if i change the theme of result. These compounds came from a docking screening. Now i just wish to test my docking predictions as true or false.  Since the substrate conc is at 100uM , i will test high scoring compounds (about 10) and some low scoring compound (about 3-4) at same conc of 100uM and the control inhibitor also at 100uM and all have same assay vol of 50 uL.

At this even molarity, If the high scoring compounds manage to get at least 50 % inhibition then isn't this data alone enough to claim that the high scoring compounds are inhibitors as compared to low scoring compounds.  Can i publish such results or IC50 is absolute must during inhibition data?

Maybe I can squeeze 1 or 2 compounds  IC50 if done in duplicate. pls advice.

Because the correlation between docking score and measured compound potency is usually quite low, a sensible approach to testing the results of docking is to try as many compounds as possible. To make this practical, you can test a single concentration. If the assay has a high signal-to-noise ratio, single measurements may be sufficient, whereas if the signal-to-noise ratio is low, averaging triplicate measurements may be needed.

Choosing the single test concentration is a trade-off. A lower test concentration reduces the amount of interference from the compound with the assay detection method (e.g. fluorescent compound in a fluorescence-based assay) and lessens the likelihood of non-specific inhibition (e.g. aggregators, denaturants, insolubility artifacts). On the other hand, a lower test concentration increases the likelihood of missing weak inhibitors. At 100 µM test concentration, you can be sure there will be assay interference and non-specific inhibition if you are testing so-called drug-like compounds. It will then be necessary to have in place methods for eliminating these undesirable hits. At 10 µM, these problems will be reduced, but not completely eliminated. This is the usual concentration for high-throughput screening of generic compound libraries. If you are screening compounds that are structurally similar to previously known potent inhibitors of the target, you may be able to use an even lower concentration.

Once you find the hits, there is still much to do to confirm that they are specific inhibitors of the target, especially if they aren't very potent. A good starting point is to measure the IC50 and the Hill slope of the IC50 curve. Single measurement is usually sufficient unless, as previously mentioned, the signal-to-noise of the assay is low. Watch out for assay interference in a plate-base assay when the inhibitor remains in the well during the readout. (see this paper: http://www.ncbi.nlm.nih.gov/pubmed/19643982). Usually, the expected Hill slope of the IC50 curve is 1, unless the iC50 is close to the enzyme concentration, or the inhibitor is time-dependent, or there are multiple cooperative inhibitor binding sites on the enzyme. If you get a steep Hill slope (>2) for a relatively weak inhibitor, there's a very high likelihood that the inhibition is non-specific.

If you have a nephelometer available, check the solubility of the compound in the assay buffer. If the IC50 is greater than or equal to the solubility, there's a good chance the inhibition is non-specific.

There are other tests for non-specific inhibition, as suggested by Brian Shoichet's group, including comparing the IC50 with and without 0.01% Triton X-100 detergent, and comparing the IC50 at low and high enzyme concentration.

Examine the purity of the sample of compound you are testing. LC-MS and NMR are good methods. It often happens that the inhibition is caused by contaminants, rather than the compound that is supposedly in the bottle.

Look for a biophysical method to demonstrate that the compound actually binds to the target. 2-D protein-observed NMR is an excellent way, but requires materials and expertise that might not be available to you. 1-D ligand-observed NMR is more accessible and requires only small amounts of protein and compound. You will still need a very good NMR instrument, such as 500 MHz with cryoprobe, and an NMR expert. ITC is great, if you have access and a lot protein. Surface plasmon resonance is also a good method, especially if there is an experienced operator to help.

These are the comments I would make as a reviewer of your manuscript if you submitted inhibition data for compounds coming out of a virtual screen. It is very easy to find inhibitors (typical high-throughput screen hit rates at 10 µM are ~1%), but not always so easy to find specific inhibitors.

Good luck with your work.

Now i am scared.  I did not know about non-specific assay interference nor of any higher and more definitive tests beyond inhibition assay. 

I will simply limit my results as docking based DPP4 hits of commercial drugs.

single to noise means blank must have very very less absorbance isn't it?

High signal-to-noise ratio is not the same as low background. The signal is the difference between the baseline and the measurement (sometimes referred to as Window). Please refer to the concept of Z-score or Z' with respect to screening, which relates the signal in the measurement to the experimental variation in the signal:

http://www.ncbi.nlm.nih.gov/pubmed/10838414

I think it is often true that people new to compound screening do not realize there can be difficulties with the results. A major aspect of careful enzyme inhibitor screening is sifting out the non-specific inhibitors, artifacts, and false positives from the small fraction of true, specific inhibitors.

The academic literature on non-specific inhibition began with this 2002 paper:

http://shoichetlab.compbio.ucsf.edu/publications/mcgovern_2002.pdf

This paper is a follow-up:

http://www.ncbi.nlm.nih.gov/pubmed/13678405

You should also familiarize yourself with the literature on the types of compounds that are widely recognized to be troublesome because they show up as hits in many screens. They have been called PAINS:

http://pubs.acs.org/doi/abs/10.1021/jm901137j

what if i use commercial assay kit? The Z factor is stated as 0.96. I didn't know what it meant until know.