There are two items: 1. the standardization of the ELISA. If quantitative, use as said before a calibration with a set of at least 5 standards and a respective calibration curve. If you add two quality controls (high and low level) you can assure that the measurement is correct. The standards can be set easily either by an international standard available via NIBSC or by a "self made standard". This can be done by the definition of a positive pool which has 100 arbitrary units. You have to make sure that your assay has a dilution linearity (one sample in different dilution has to give the same final result: i.e. by using a serial dilution). Beware: in the detection of specific antibodies it’s not that easy to reach because of the fact that two properties influencing the ELISA signal: the antibody concentration and the affinity. The affinity will influence the slope of the curve. Both parameters are highly individual. So you have to make sure, that the concentration is the only parameter influencing the ELISA signal. For this purpose I suggest to use as incubation buffer: 0.01 mol/L phosphate pH 7.2, 0.3 mol/L NaCl, 3% v/v Gelafusal (digested gelatin with normally no cross reactions etc), 0.1% v/v tween 20, 0.001 g/L phenol red.

Second: The normalization as term is used to get the ELISA results from varying surrounding conditions into correlation with parameters you are trying to influence. So you can use the total protein content, the concentration of housekeeping proteins (SOD, LDH, …) whatever is constant per cell number etc. It’s can be also the number of cells, positive cells (in FACS, Histology, …).

For normalised your sample, you should use a control sample in all the experiments. This control will be treat as standard ELISA unit (i,e 100) for all plate. After that you have to calculate the ELISA unit from the OD value. Suppose the OD value of your control is 1.54. That is 100 , other samples will be 100/1.54 X OD of the unknown sample, the value will be ELISA unit. You should remember that the control should be taken from equal volume of 7-8 sample of different range (OD). 

Hi Ahad,

Here, I use serum as primary antibody. It contains number of proteins. This is what I want to normalize. Can I prepare a standard for serum? Content of serum is different from person to person. isn't it a problem?

Yes. I'm looking for leishmania antibodies in serum against parasites. In every assay I have done, high background was observed for normal serum than known positives. That's why I think It may be due to the differences within serum samples person to person. Even we get a same sample volume, the protein content may differ. I'm using indirect ELISA. According literature, the high background may be resulted due to non-specific bindings with other proteins. But in positive sera, I have never seen such a matter. That is because of specific antibodies they have. In negative serum, we can't see target antibodies. Therefore, other serum proteins are binding non-specifically. First thing, I should do is reduce non-specific bindings. Am I correct?

Can I use the OD value of each serum samples without adding coating antigen as the internal control? It gives some idea about the content of total protein? But the problem is, we don't know the limit of maximum amount which can coat with wells. In sera, which have higher protein content, the OD value may be same. Therefore I'm thinking of micro-Lowry assay to quantify the protein of each sera. any suggestions??

There are two items: 1. the standardization of the ELISA. If quantitative, use as said before a calibration with a set of at least 5 standards and a respective calibration curve. If you add two quality controls (high and low level) you can assure that the measurement is correct. The standards can be set easily either by an international standard available via NIBSC or by a "self made standard". This can be done by the definition of a positive pool which has 100 arbitrary units. You have to make sure that your assay has a dilution linearity (one sample in different dilution has to give the same final result: i.e. by using a serial dilution). Beware: in the detection of specific antibodies it’s not that easy to reach because of the fact that two properties influencing the ELISA signal: the antibody concentration and the affinity. The affinity will influence the slope of the curve. Both parameters are highly individual. So you have to make sure, that the concentration is the only parameter influencing the ELISA signal. For this purpose I suggest to use as incubation buffer: 0.01 mol/L phosphate pH 7.2, 0.3 mol/L NaCl, 3% v/v Gelafusal (digested gelatin with normally no cross reactions etc), 0.1% v/v tween 20, 0.001 g/L phenol red.

Second: The normalization as term is used to get the ELISA results from varying surrounding conditions into correlation with parameters you are trying to influence. So you can use the total protein content, the concentration of housekeeping proteins (SOD, LDH, …) whatever is constant per cell number etc. It’s can be also the number of cells, positive cells (in FACS, Histology, …).

Thank you for all valuable answers. My problem was, I got high OD values for normal sera than positive sera. Now, I have some idea to resolve this.

1. Main thing is analysis part. we can't just obtain OD for comparison. We have to normalize each OD with an internal control (eg. we can do micro-Lowry assay for each sera to calculate the total protein content) because the surrounding conditions may differ within each sera.

2. If we get high OD values for normal sera even after normalize the OD values, almost it may due to non-specific binding. We have to reduce the non-specific bindings. There are several ways, we can increase NaCl concentration in PBS, can reduce the incubation time of sera, can reduce the conjugate concentration, increase the efficiency of coating antigen and etc.

Now you have given the real point. Your normal OD is greeter than positive control....Second point non specific binding could be interfering. These two are making problem for you. 

First you clean up your ELISA by blocking the plate with either BSA or Casein or skim milk and be sure that the plate is blocked for non specific binding.

Second part, the OD of negative serum is higher . The presence of antibody depend on the total protein of the individual serum. So you calculate the total serum protein and measure antibody. You will get a ratio of both positive and negative putting antibody(OD value) /protein. You calculate all individual serum like this and compair their ratio with positive and negative ratio value.

You cannot determine level of antibodies from a single dilution.  You should test serial dilutions of both positive and negative samples and look for the difference in the curve.  Initially use 1:10 dilution steps.  Then, when you will know at which dilution each sample does not result in a signal above background, you can use more appropriate dilutions steps.

If you see that the negative samples generate higher signal than the positive you should do either  or both:

1. Use higher dilutions of the samples;

2. Dilute all samples in a serum, either animal or human.  Not full serum, of course, but 10 - 20% in PBS + 0.1 - 0.5 detergent.

Dear Falk,

you can determine the antibody concentration using a single dilution only. You have to make sure only that you have a dilution linearity. So you have to adjust your test in a way, that only the concentration correlates with the signal and not the affinity or the composition of both, affinity and concentration. It's can be done by using a sample dilution buffer with higher salt concentrations (0.3 ... 0.5 mol/L NaCl), with increased inert protein concentraions or with a modified test design (i.e. with an anti-Ig or the antigen at the solid phase and an labelled antigen as conjugate: see attached paper). 

As a general rule one measures the titer of an antibody, not the concentration.  The titer is a function of the mixed concentrations and  affinity constants of polyclonal antibodies.  The titer is expressed as the dilution at which binding can just be detected.  Serial dilution of the sample in a standard negative serum sample  and determination of binding to the antigen is all that is required.  If a labeled anti-globulin antibody is used a standard sample can be used that does not bind to the anti-globulin, for example animal or synthetic sera.  Kiessig's method only works if the antibody concentration is much higher than its dissociation constant. His alternate approach requires complete removal of immunogloblins from the serum and determination of the binding sites of the captured proteins using a labeled antigen. Because mixed antibody affinities can result in concentration dependent binding of labeled antigen the result may be no better than the much simpler titer determination. 

Dear 

you can also try to Use of the Normalized Absorbance Ratio as an Internal Standardization Approach to Minimize Measurement Error in ELISA