Regarding Linearity Range

[aniket]

Dear Forum,

I am working for a Discovery DMPK lab, my main job is to screen to many analyse many compounds in week in biological samples, my main problem is i am not knowing the expected unknown concentrations because this first time the molecule is being studied. so i use a generic range of calibration stanards for quantifing my unknowns, so can you suggest me some range of calibration range??

Should i use an Internal Stanard for my analysis, i am doing scerrning so some of my friends suggested not to use is, is it a good strategy for bioanalysis? i am bit skeptical analysis without IS!

Can you please guide me on this?

[danko]

Hi aniket,

I assume that - by bioanalysis - you mean peptides/proteins.
If your detection wavelength is 210 – 220 nm, the absorbance intensity depends on the number of peptide bonds. So if you have an idea of how many amino acids your target molecule consists of, then you can calibrate with a known peptide/protein and correct the value with the difference between the number of amino acids.
Of course this will be a very, very rough estimation, but maybe that is all you need at this stage.

If you haven’t got an idea of the size of the peptide of your interest, you can get even more rough estimation of that by separating it (together with some known peptides/proteins) on SEC/GPC.

Yes, many rough estimations, but what do you do, if MS is not available in your lab?

Good luck

[aniket]

dear forum
thank you for your replay. Bioanalysis meance i have to analyse the known NCE by spiking it in differant matrices such as rat, mice, human plasma or urine etc....so please consider the same and guide me...

[DR]

You'll have to see what the anticipated dosing range is (max mg/Kg), look up what % blood, by weight, most mammals are and go with that as your high end approximation. Then keep making dilutions until you can no longer quantitate accurately, and until you can no longer see it (LOQ, LOD).

Disclaimer:
I haven't done these myself, but it's how I'd approach it unless offered a better way.

[danko]

Hi aniket,

Obviously I have misunderstood completely your initial question. Sorry.

I would go for spiking the various matrices with the known drug/standard, as long as the method of analysis is not validated for the different matrices. Especially resolution and sensitivity are important to document before concluding that the matrix does not affect your results.

I hope this time I got your question right. Or is it just getting funnier and funnier?

Best Regards

[Mark Tracy]

Internal standards can't fix problems of co-eluting matrix components, and with an unknown matrix, the IS is also subject to matrix interference, so now you have two problems to worry about. The IS (or more properly the surrogate) may be useful to correct for inaccurate measurements in the sample prep or correct for inefficient extraction, but only if it behaves substantially the same as the target analyte. Also, IS degrades the LOD and LOQ by 40% under the best of circumstances.

As a historical note, the internal standard became common practice in the early days when detector response factors tended to drift badly, and manual injections were of dubious repeatability. Most modern detectors and autosamplers remove the need for IS to compensate for them. One important exception: MS still benefits from IS because of response drift; here isotopically labeled IS is the best.

Bottom line: use IS when there is a demonstrated need for it, otherwise don't.

[aniket]

Dear Members,

Thank you all for replying. Danko this time you understood my question correctlly. This is what i wanted to say.

One more thing i would like to clarify is suppose my unknown concentrations are 2-4 times higher than ULOQ in that case what should i do? Should i dilute the samples(without proving dilution integrity tests) or should i explore using an higher ULOQ(but i assume there are some practical limitation to this), so what is the best approach??

Dear mark, i am confused by your stance on IS? Many methods for NCE reported in litreature contain IS? Can you please explain in detail about without is methods?? i am using LLE, SPE for sample claenup.

Ok thats is a pretty big question!!

Take care,
Aniket

[Bruce Hamilton]

Dear Members,

One more thing i would like to clarify is suppose my unknown concentrations are 2-4 times higher than ULOQ in that case what should i do? Should i dilute the samples(without proving dilution integrity tests) or should i explore using an higher ULOQ(but i assume there are some practical limitation to this), so what is the best approach??
Aniket

It's really up to what data quality your clients expect. I would redetermine the ULOQ if there were a significant number of samples that are above the existing ULOQ. However, there are some obvious concerns...

I would have based the original ULOQ on the expected concentrations in samples, and if samples are much higher than expected, you should be talking to your clients before working on the protocol.

You have to ensure that higher quantities of analytes in samples are still within the original method sample preparation qualification data. It could be that higher concentrations of analyte could also invalidate your previous sample preparation protocols.

It's a bad idea to add a dilution step to some samples, and not to others, but few samples are probably best handled that way, provided you also demonstrate the rest of the procedure can handle the extra analyte.

With regard to the use of an Internal Standard, then, once again, your clients should be specifying whether one should be used. I'd want them to sign off on the agreed protocol, based one what they consider " best practice ".

Please keep having fun,

Bruce Hamilton

[DR]

Use of an IS is most appropriate when your sample work-up contains extraction steps that may, or may not be very efficient. Presumably, you choose an IS so that any losses due to extraction steps will be about the same for the IS as for the analyte. This way, when you have 4 different people shaking sep. funnels or using SPE columns, you don't get 4 different answers for the same sample.

[Mark Tracy]

In your case using LLE and SPE, an IS might be a good idea, and certainly should be investigated. But before you can rely on it, you must have some demonstration that the IS behaves the same as the target analytes in the sample prep. That is the absolute recovery of target and IS is the same over the concentration range of interest. Only then can you use the ratio of peak areas as a correction. Second, you must establish that the matrix does not cause a problem. Matrix problems could be selective binding/precipitation of either the target or the IS. Another matrix problem could be that there is an interfering peak under either the target or IS, and that would invalidate the assumptions of IS calculations.

When you are analyzing unknowns in variable matrices the IS simply doubles the number of problems that you must control.

If you can establish consistent recovery >90% without IS, you probably don't need IS. There are some situations where you need IS: working with very small samples where quantitative transfer is not possible; the target recoveries are variable; the target recoveries are low; the detector response is variable; the sample injection volume is not reliable.

There is nothing preventing you from calculating your results by external standarization and internal standardization from the same data set. If IS gives demonstrably better results, use it, otherwise don't.

[HW Mueller]

This sentence by Mark:
"There is nothing preventing you from calculating your results by external standarization and internal standardization from the same data set.",

straightens it all out.

The interpretation which I have given several times is: Use an internal standard all the time if you can find a good one, you can use it as a semiquantitative check if you don´t want to base calcs on it, etc.

[AdrianF]

As previously mentioned there are many problems associated with internal standards

You can avoid using them by spiking the matrix eg serum with the molecule of interest. The sample and standards are then extracted in exactly the same way- recovery is then automaticaly corrected.

If samples are above ULOQ then dilute them into the range using the same matrix.

I am interested at how you arrived at the ULOQ.

[HW Mueller]

AdrianF, what do you do if you can´t get an analyte free matrix?

[AdrianF]

In Reply to H W M
As new chemical entities(NCE) are being discussed here the problem will not arise.

However if analyte was present in the matrix and I wanted to avoid the use of an internal standard, I would pool enough for the experiment and then calculate the amount present by means of standard additions. By trial and error I would add the chemical of interest to about double the signal from the molecule and hence calculate the amount in the matrix which could then be allowed for.

[HW Mueller]

And that´s better than external + internal standard?