Worst method write-ups

Off-topic conversations and chit-chat.

11 posts Page 1 of 1
This could almost be a name-and-shame thread, except I don't want to get anyone into personal arguments or bring ChromForum into difficulties with irate authors!
But does anyone else find it frustrating that methods sections in publications often appear to be written by someone who either had no idea what the method was, or no inclination to share it?

I've just been asked, quite reasonably, to reproduce an HPLC analysis where the original method tells me how the sample was made (good) what instrument was used (not especially helpful), that the column was a core-shell from Phenomenex, with its physical dimensions. It neglects to tell me the column chemistry, the solvents, or the gradient (i.e. everything that actually mattered). It ends by saying "Internal standards for <the names of all the analytes, none being an internal standard> were used to generate standard curves for quantification of the respective compounds." which leaves me thoroughly puzzled: did they use internal standard calibration or external?

There's a reviewing issue too: it's quite difficult for 2 or 3 reviewers to have expertise in every part of a multi-disciplinary paper, and it's quite likely 1 or 2 of the reviewers are sufficiently senior that they no longer have any contact with lab-work. Being unpaid and busy, the reviewers probably won't try too hard anyway, and if the methods are buried in some supplementary stuff in a different place to the main paper, it's quite likely some reviewers won't even read it. Clearly no one who knew anything about HPLC read this method.

This sort of situation seems to happen so depressingly regularly.
Do you mean those papers where the person who wrote it up certainly didn't do the lab work and the person who did is long gone? Seems to be more common where one group puts out a lot of papers.
I had one synthesis where if you followed the procedure as written it didn't work and when we called the author he admitted that they had not done it that way, they had used a short cut to save time. The short cut worked fine but the as written procedure degraded the starting materials.
Working as I do on the ragged edge between animal behaviour and chemistry I see a lot of papers with critical bits left out and irrelevant stuff included. Usually it is because the biologists took some samples to some chemists and said analyse these. Fortunately there are not so many variables in GC-MS as in HPLC, and the methods are pretty generic so it would be straightforward to replicate them.

Peter
Peter Apps
It just irritates me that people who'd never consider falsifying or mis-recording the results have no worries about falsifying or mis-recording the method.
The methods are what generates the data. If the method is misleading, the author might just as well have used Excel's random-number generator to make the results, or swapped all his chemical names around at random.
(sorry: you'll realise that I feel strongly about the experimental side of things. A story only lasts until the next better story; interpretations of results change as time goes along. Data are for ever. A good measurement remains the best estimate for underpinning whatever hypothesis is currently favoured, until such time as someone shows it to be wrong, and/or does a better job of measuring it.)
That's a good point, and looked at from the other direction; if they claim to have used a method that could not have worked (or they apparently know so little about it that they omit critical parameters), then how good are the results ?.

Similarly; if I have to referee a paper and it is full of spelling errors, sloppy logic and disorganized presentation then I conclude that they probably were not very careful in the lab either and the results are not worth publishing.

Peter
Peter Apps
Peter,
I have seen the same thing in patents, I believe sometimes the person doing the lab work is not the same person that edits the document for publication. Hence "non-vital" details are missed off. I had to laugh at the comment "working at the ragged edge..." I sometimes find myself in the same position between sanity and complete madness!

GCguy
GCguy
A person from an industrial background once told me, quite seriously, that the art of writing patents was to ensure that you put enough stuff in it to pull the wool over the patent examiner's eyes so that he thinks it is a complete method, while leaving out enough to ensure that the patent doesn't actually work. This prevents your competitors from making use of it when it expires, and makes it possible to protect an idea for longer, by patenting further details at a later date.

At least he was being honest.
In the consumer products industry, the 1970s patent for solid antiperspirants was written inadequately, and other companies were able to go outside the patent range for stearyl alcohol content and skirt the patent. So that was an error costing quite a few million dollars. Think of what a royalty of 1/2 cent per unit would've been worth....
Then there are the companies that create snowstorms of patents on marginally relevant stuff to conceal what they are really up to.

Peter
Peter Apps
Peter,

Your example reminds me of an example of a method we received from a client for doing testing by IC on a product they make. Their in house chemist developed the method, and when I read through it I knew it could not work, yet they always had passing QC when it was analyzed in house.

The method stated that the calibration curve was to be analyzed at levels of 0.1, 1.0 and 2.0 ppm and the QC check standard was to be made at 5.0ppm(this was the first oddity I saw, doing calibration check with standard far above the upper limit of the calibration curve). After reviewing the procedure for preparing the QC check solution I saw they were taking 10ml of a 100ppm stock standard and diluting to 100ml final volume for a final concentration of 5ppm. They did this every day and the QC check sample always quantified near 5ppm. That sounds like magic to me :) but it always passed on their records.

Needless to say we extended the calibration with 5ppm and 10ppm standards and made a proper 5ppm check solution which also passed when we ran it.
The past is there to guide us into the future, not to dwell in.
James, somewhere in that customers QC lab was a sacred volumetric flask that was handed down from generation to generation of analysts because it was the only one that made check standards that passed QC. It was so old that the numbers had rubbed off, and it had to be kept separate from the other flasks in case it got mixed up with the other 5 ml flasks that were the same size, or the other 10 ml flasks that were the specified volume but didn't work properly.

Let us hope that we all make it through 2018 without being poisoned by improperly formulated products !

Peter
Peter Apps
11 posts Page 1 of 1

Who is online

In total there are 5 users online :: 0 registered, 0 hidden and 5 guests (based on users active over the past 5 minutes)
Most users ever online was 209 on Wed Nov 29, 2017 10:42 am

Users browsing this forum: No registered users and 5 guests