It seems like the first step might be to make sure that you have calculated your primer efficiency for each set of primers you are using. Just like with a loading control gene, primer efficiency is used to equalize between different genes

You can determine your primer efficiency by performing qRT on a serial dilution of a known concentration of a copy of your gene of interest that has been cloned into some kind of easy cloning vector (such as pGEMT or TOPO TA).

However, there are some programs that allow you to calculate your primer and PCR efficiency without performing such a time-consuming step (of cloning and then serial dilution qRT). See the attached manuscript for one such program that I have used in the past.

As an aside, how do you know that some of your targets should be of higher abundance, since it appears you are looking at DNA from a mixed population?

Following the question of Andrew Nelson, I assume you are currently testing the method by using other quantification methods in parallel?

To estimate efficiencies for each well without the need of standard curves (especially if the qPCR are already done !) I would also recommend LinReg which is quite simple to use but flexible enough to correct an error in one reading. A wider comparison of methods and software has been published by Ruijter et al.

noble.org/Global/CoreFacilities/Genomics/LinRegPCR_ramakers2003.pdf

Article Evaluation of qPCR curve analysis methods for reliable bioma...

Reconciling the two universes of a purified standard and a biological sample extract - aiming to amplify the same target in both contexts, can be done using the attached relationship.

I don't understand your problem: you saiy you used a "standard curve from which the target quantities were calculated absolutely" and later you complain that "the Ct values are not in agreement with the quantities"...

The Ct values are interpretable only relative to the corresponding standard curve, giving you a result (concentration or amount) relative to the standard (and if the amount or concentration of the standard is known absolutely, this is called an "absolute" result). So the Ct values are eventually not interesting.

The methods only works when the standard amplifies with the same efficiency as the samples. If these efficiencies are different, the results are biased. In such cases it can happen that in a mix of more A and less B you eventually "measure" more B and less A (is this what you were saying you observed?)

Efficiencies can be different between standards and samples due to the template structure (high-molecular vs. low-molecular DNA) and differences in the presence of contaminating PCR inhibitors (e.g. residual phenol from TRIZOL purified standards that is not present in the samples that were purified over columns... or residual huminic acids from soil samples from which the bacteria were isolated... or whatever). Some different residing substances can modify the amplification curves(profiles) and affect the background-correction, detrending and other data-processing steps, so that systematically biased Ct values can be the result. A combination of both can be the case...

It is sometimes difficult to find/identify such differences. In any case you should prepare a dilution series of your sample and see if the estimates for the amplification efficiency are similar to those of the standards.

Ideally (but not always possible) standards and samples should contain the target sequence in molecules of similar complexity and the DNA should be in the very same buffer. Preparing a good standard can be a painful job.