Dilution factor in HPLC analysis

[adamos1945]

hello
I would like to ask for help regarding calculation of dilution factor to get the amount of mycotoxin per gram of sample. the extraction protocol as fellow:
-ground 5 g of sample with 25 mL of methanol: ACN: water (25:25:50).
-shake for 20 min and centrifuge for 10 min at 2500 xg.
-filter supernatant through filter paper.
-take 10 mL of extract, mix with 40 mL of PBS and filter through microfiber filter.
-pass 10 mL of diluted extract through Vicam column, wash column using 10 mL of PBS, and elute fumonisins by passing 1.5 mL of HPLC grade methanol. collect the elute in vial.
-dry down methanol elute.
-redissolve elute in 400 uL of ACN: water (50:50).
-mix 30 uL of well-mixed redissolved elute with 30 uL OPA reagent.
-inject 30 uL.
we made the standard curve using ppm as unit. so for calculating the concentration , should we use ppm as the standard curve?
how to calculate the dilution factor for this protocol to get amount of mycotoxin per g of sample? also, we used ppm for making the standard curve, so while calculating the amount of mycotoxin per g of sample, should i use the results obtain through standard curve or should i converted to concentration as ug/ml?
thanks for your help.

[James_Ball]

If you use ppm your units will not cancel when checking the calculation so it is more difficult to know if you have it correct.

While a dilution factor can be calculated, we need to know how you prepared your standards to know how it will be applied to the final instrument reading before we can give a correct answer. For example if you simply dilute an amount of analyte into solvent in inject then a direct application of the dilution factor works, if you process the standard through all or just a portion of the preparation process then it must be applied differently.

[adamos1945]

If you use ppm your units will not cancel when checking the calculation so it is more difficult to know if you have it correct.

While a dilution factor can be calculated, we need to know how you prepared your standards to know how it will be applied to the final instrument reading before we can give a correct answer. For example if you simply dilute an amount of analyte into solvent in inject then a direct application of the dilution factor works, if you process the standard through all or just a portion of the preparation process then it must be applied differently.

thank you for your answer.
for preparation of standard, we prepare the desired concentration such as 0.5 ppm, 1 ppm, 5 ppm and 10 ppm into vial, then we dry them and redissolve them using 400 uL of ACN: water (50:50), then mix 30 uL of well-mixed redissolved elute with 30 uL OPA reagent, finally inject 30 ul into HPLC.

[Arne]

I think this rather complicated sample preparation procedure calls for an internal standard - ideally an isotopically labeled analyte or something substantially similar that is added at the very beginning.

[lmh]

This is probably the most important post you will ever read in your life:

With an extraction procedure like that, the way to calculate the dilution factors is this:
Draw a little diagram of the entire extraction procedure. Draw little tubes, and arrows pointing from one stage to the next, showing what volume you took out, what volume you put in, where everything is.
You can work one way or the other. You can either start at the beginning with 5 gr of sample, or you can start at the end with a final result of X units fumonisin (more about units later!).
If you start at the beginning, work out how much original ground material corresponds to each solution you make, all the way through. You start with 5g extracted to 25mL. You dilute it in all sorts of ways, and add other volumes, and I think (but please check) that by the end you have 60uL of derivatised solution equivalent to 0.03g of your original ground material. You inject 30uL of this, so you're injection came from 0.015g of ground material.
Let's assume you do your calibration curve in mg/mL. If your result from injecting a sample (that contained fumonisins from 0.015g ground material) came out at X mg/mL, we know that you had 0.030X mg of fumonisin derived from 15 mg ground material, so it's a fairly simple matter to work out the ppm by weight (0.030X / 15 * a million)

If you start at the end, with a result of X units of fumonisin, you'll find it much easier if the units are something easy to handle. I don't recommend ppm because it gets really mind-twisting when you have a dried-down stage in there! If your calibration curve is in mg/mL fumonisin, you can easily work backwards by working out how many mg (ug, ng...) fumonisin the original 30uL injection contained... so how much did the entire 60uL derivatisation reaction contain? (twice as much), so how much did the 400uL redissolved dried stuff contain?... so how much was sitting in the dry pellet... how much was eluted from the Vicam column... etc. etc. right back to how much, in mg, was present in the original 5g of material that you extracted.

Do dilution calculations this way, and you will rarely, if ever, get them wrong.
(Now I'm going to look a right wally when someone points out that I have indeed made a mistake, but my excuse is I did this really quickly on the back of an envelope that wasn't quite big enough).

(but further to the last comment, yes, there are lots of ways to lose the analyte in an extraction like this, so an internal standard is a good idea. If you don't have access to an internal standard, at least run spiked experiments where you extract material in parallel, some of it spiked with the authentic product, and then check that at the end of the extraction, you get back what you expected. If your recovery is 90-110% you can be very happy. If it's 15% you probably need to panic!

[Arne]

Great advice and explanation to go step-by-step, lmh, regardless of whether the math was exactly right or not! it is also good to actually draw things.

Substituting a spike-recovery experiment for an internal standard is also sound advice, assuming that recovery is constant.

Repeating the spiking for one sample out of each series of similar samples would also be judicious.

[Hollow]

With an extraction procedure like that, the way to calculate the dilution factors is this:
Draw a little diagram of the entire extraction procedure. Draw little tubes, and arrows pointing from one stage to the next, showing what volume you took out, what volume you put in, where everything is.

It's nice to see someone else is advertising this visual procedure as well.
And after that, before going to the spreadsheet, it's good to write down the algebraic expressions for each step, then connect them to one big expression.
Like this, you know what cells you'll need in your spreadsheet and how to connect them (every parameter (raw data) in the expression gets it's own cell).

But as with every advertising, some people seem to be immune against them, even against the good ones...

[James_Ball]

Great advice and explanation to go step-by-step, lmh, regardless of whether the math was exactly right or not! it is also good to actually draw things.

Substituting a spike-recovery experiment for an internal standard is also sound advice, assuming that recovery is constant.

Repeating the spiking for one sample out of each series of similar samples would also be judicious.

I agree with it also. One of the reasons I asked how the standard was prepared, because I have seen some run a standard through the extraction process which nullifies any calculations, but then do the calculations and double correct the final value. Best if you can include the units at each step and if they all cancel to give a final result unit that makes sense it is another check of the math. If you end up with something like mg/g/ml, then something went wrong.