Do research has papers lie? Problems recreating methods

[Chromatographyrocks]

Hi guys first post! I just got my hands on a brand new Xevo TQ-S with an I class aquity system! First lcmsms in the entire company! (no pressure) My question is can research papers be wrong or even intentionally lie? Here's my scenario:

Project: Amoxicillin and Clavulanic Acid bioanalysis in canine/feline plasmas using LCMSMS

I've tried to recreate a method for this analysis from multiple papers however there's something bugging me so much I had to ask for opinions. The paper uses a C18 BEH column 50mm 1.7um and mobile phase of 85:15 ACN:2mM AmmForm. From my experience of working with this compounds in RP on HPLC I know to retain on said column we need to work around 5% organic or else there is zero retention. I know this but tried it anyway on the ms and my suspicions were confirmed! The authors claim RTs of 1.5mins! The thing is pretty much every other paper I can get my hands on all use very high % organic! I just can't get my head around it! I have other issues too specially with the sensitivity for Clavulanic acid I can't get low enough! So I have another q! Does polarity switching affect the sensitivity of some compounds? I'm running amoxicillin in pos and clavulanic acid in neg.

Im stumped, I'm currently running on a HILIC column now with good retention but still suffering to get to 1ugml for clav. Any advice and discussion is appreciated

Thanks

Oh here is the main paper I'm going off
https://www.google.co.uk/url?sa=t&sourc ... Xkt7DI2ceA

[tom jupille]

Uhhh, I'm missing something here. As far as I can see, that paper used a COSMOSIL 5C184PAQ column and the mobile phase was, and I quote: "0.2% formic acid in water–0.2% formic acid in methanol (20:80, v/v) and set at a flow rate of 0.6 mL min"
.

[Chromatographyrocks]

My apologies wrong link, I have a few papers here is the correct paper:

https://www.google.co.uk/url?sa=t&sourc ... 4XUdNXMKnw

Even in that paper at 80% methanol I don't believe there would be any retention! Both amoxicillin and clavulanic acid are extremely polar, hence why I am getting good results using using HILIC. As I said above I have worked with both compounds many times on hplc and we always use 5% or even 2% organic! And it's the reason why we got the lcmsms. How can so many papers say this yet I can't replicate it! Very confused

[lmh]

Research papers only rarely lie deliberately, but lots of things can make it hard to reproduce something from a paper. I'm not for a moment suggesting that any of the following apply to the method and publications you're looking at, but things I've suspected or met in the past:

(1) Methods that only work because the column has had a long and interesting history. Some methods rely on the chromatographic characteristics of the junk that has become permanently bound on the column from previous samples, or on interactions with silanols that were once end-capped, but whose end-capping has long since vanished.
(2) Methods where the bottles got reversed between when the samples were run and when the paper got written.
(3) Methods where there are two parallel written records: the trail of people saying "this was measured according to X" because that's what the last publication said, and meanwhile the increasingly scruffy photocopy from someone's lab-book that says what you should really do, and is handed down from post-doc to student.
(4) Methods that only work because the original hardware wasn't doing what it said it was.
(5) Methods that look like they worked but were never working (e.g. all those where there is actually no retention, had anyone checked).

There are, of course, some genuine situations were very strange things happen. For example, there are analytes that have U-shaped curves of mobility versus solvent strength. These will work quite nicely at low solvent strength, or high, but not in between, and in gradient methods they may come out in the run after the one where they were injected, despite an extensive column wash and good re-equilibration.

[Chromatographyrocks]

Thanks for your extensive answer! I had considered some these possibilities and even the possibility of copy and paste errors. However there are at least 5 of the papers use this high organic and the paper linked above has summarised these in a nice table just still doesn't compute with me, I can understand one or two instances of this occuring but 5 or 6 could rule out more than one of your suggestions (but I like the out of the box thinking!) I think, like you said using a column with a colourful history is quite likely I suppose. I hadn't even thought about the U shaped mobility but I did try to replicate it and the compounds shot of the column so fast it left me reeling!

Thanks again for the discussion. I have retention on HILIC so it's not all bad, and I had been looking for an excuse to use it haha. Do you have any thoughts on polarity switching for the ms? I can understand in older systems with long switching times it may be an issue but with the latest high spec equipment with 100ms or less switching I wouldn't have thought it to be an issue. The problem with clavulanic acid is it just falls apart with only 2kw on the esi whereas amoxicillin requires 3-4. Any tricks of the trade for improving sensitivity? I need to gain around 10times more. I'm down to 5ug/ml (injecting 10ul) but need to go to 1 (which is no problem for amox)

Thanks again

[mckrause]

We have seen similar situations in multiresidue pesticide analysis, where papers claim outrageous recoveries and precisions (98.9-100.6% jumps to mind). I don't think they lie, but you have to remember that many times papers are being published by people with extremely little analytical experience and very limited skill sets. This can certainly lead to misinterpretation of data.

Unfortunately this situation is not limited to published papers. We have tried our best to replicate some of the published LC-MS/MS chromatograms for multiresidue pesticides without much success. We can separate the compounds and find them, but the published elution times and orders do not jive with our data even when we are using the specific column and analytical conditions.

In your particular case I would agree that high organic would not seem to be correct. Our experience mirrors yours, where low organic with "aqueous" phases or HILIC seem to be called for.

[MSCHemist]

I'm banging my head with a series of papers by the same author using trifluoroacetylacetone and ethyl chloroformate to do amino acids. The FID chromatograms look beautiful. However the more I inspect the paper the more problems I find. I've come to the conclusion that the author is mistaken that his unique derivatives are ethyl carbamates from the ethyl chloroformate and methyl esters from the methanol. The biggest evidence is that acetylacetone reacts with primary amines yet proline is right in line with the other amino acids.

I've run into quite a few papers that look very promising but turn out to duds.