Every letter we have printed

The most useful sentence in the piece is the one saying a bad document is not a bad product. I have spent two years on forums watching people conclude the opposite from a missing signature block, and it has made the whole conversation about honesty rather than about paperwork.

Corrections are the missing practice. Every publication has a mechanism for saying it got something wrong; almost no testing arrangement in this trade does. A reissued certificate with a note explaining what changed would be an ordinary professional courtesy and I have seen it twice.

What proportion of participants reached the target dose is a single number reported in most of these papers and quoted in almost none. It is the most direct statement of tolerability the study contains.

Grading scales designed for oncology are used in this literature and they fit poorly. A scale calibrated against life-threatening events compresses everything a person in these studies is likely to experience into its lowest two categories.

Meal composition interacts with the mechanism in ways the literature has measured and popular coverage never mentions. Fat content in particular changes emptying substantially at baseline, before anything else is introduced.

Reading the substudy protocols rather than the papers is instructive: the scan schedule, the fasting requirement and the positioning instructions are all specified and all affect the result. None of it appears in secondary coverage.

Publishing the individual-level scatter rather than the group means would transform this literature. The means are compatible with several very different underlying distributions and the scatter would distinguish them immediately.

I train four times a week, eat a hundred and sixty grams of protein and my appendicular lean mass has fallen by 1.8 kg over ten months while every lift has gone up. Your section on mass against function was the first thing I have read that made that seem normal rather than a failure.

Your piece treats the one-quarter rule as discredited and then quotes fractions of one third and two fifths from the substudies as though those were more solid. They are group means from a hundred and forty people. Physician, heal thyself.

Regional and total measurements behave differently and are reported interchangeably. A total figure can be stable while a regional one moves substantially, and both are described as bone density.

Reasons recorded at the point of sale would be a different dataset entirely and one nobody has ever assembled. A supplier knows precisely who stopped ordering and when, and has no reason at all to publish it.

Reasons cluster by phase. Early discontinuation is dominated by tolerability, late discontinuation by cost and circumstance, and reporting them as a single distribution loses the only structure in the data.

A publication that covers a supply chain has something to say about maintenance that a clinical publication cannot: the arrangement depends on a company continuing to exist, continuing to make the same thing and continuing to ship it. None of those is guaranteed over a decade.

Research-grade material is not sold as sterile, is not claimed to be sterile and is not approved for human use in any jurisdiction. That is a straightforward statement of what the grade is, and your department is one of the few places in this trade where it is made without embarrassment.

The reason nobody discusses this is partly that discussing it invites a use-case that the research-use framing exists to avoid. The silence is structural rather than evasive.

Reporting the negative results is what would distinguish a serious programme. Every honest series contains determinations that found nothing interesting, and a publication history composed entirely of findings is a publication history that has been curated.

Retained samples are the part I would add. Half the vial tested and half sealed and stored means a disputed result can be re-examined a year later, and it costs a freezer shelf. Almost no programme does it.

A suggestion for the services rather than for you: publish a price for a full package including the raw data files. I would pay it, my colleagues would pay it, and at present there is no line on any of the four price lists that lets me.

What a badge cannot do is carry forward in time. A report on a lot filled eighteen months ago says nothing about the lot shipping today, and the badge sits in the page template rather than on the product. Verification has a date and the design of these things hides it.

A brief thank you from a reader who works well outside this field and follows it anyway. The writing assumes intelligence without assuming background, which is harder than it looks.

Where two activities produce the same effect by different routes, the combined result may be less than the sum, and that is the direction the pharmacology usually predicts. Popular accounts assume the other direction without argument.

The activity ratio is a design parameter chosen by the chemists, and different programmes chose differently for reasons that are published. Reading those design rationales is more informative than any comparison of outcomes.

Species-matched receptor is the minimum condition for a potency comparison to be meaningful, and several widely circulated tables mix figures obtained on receptors from different organisms. The mixing is not visible in the table.

Albumin binding modifies the free fraction and therefore what actually reaches the receptor, and it differs substantially between the compounds under discussion. Potency comparisons that ignore it are comparing the wrong quantity.

The half-life to steady-state arithmetic deserves to be spelled out every time, because it is the single most useful calculation in this subject. Four to five half-lives to approach steady state means a weekly molecule with a week-long half-life takes over a month, and every dose-response impression formed before then is formed on incomplete exposure.

Deconvolution software is doing more work than most readers realise. It will happily produce a confident neutral mass from an envelope with two overlapping series, and the output is a single clean number with no indication that it was a judgement. Ask for the raw envelope, not the deconvoluted figure.

A small point on units. Parts per million mass accuracy is a proportion, so five parts per million on a small peptide is a few thousandths of a dalton and on a large one is rather more. Quoting an accuracy specification without the mass it applies to is the same error your chromatography pieces complain about.

I would add one omission to your six lines: the date and nature of the last calibration. A parts-per-million figure from an instrument last calibrated a fortnight ago is a different claim from one calibrated that morning with an internal standard.

The invisible category that worries me most is the residual solvent, because it has no chromophore, ionises badly and is present by design rather than by accident. It is measured by a completely different instrument and almost never ordered.

Aseptic technique is a set of behaviours with a defined purpose and it is described in this market as a list of products to buy. The purpose is to avoid introducing organisms at a defined set of moments, and knowing which moments matters more than any consumable.

The stopper is the critical surface and the one people touch. Once a stopper has been handled, the disinfection performed before handling it is no longer relevant, and the order of operations is the whole technique.

Angle follows length rather than preference. A short needle goes in perpendicular; a longer one into a pinched fold at an angle. Treating the two as interchangeable styles rather than as consequences of the equipment is where most of the confusion sits.

A practical point about drawing up. A needle wide enough to withdraw viscous solution quickly is not the needle anybody wants for the injection itself, and swapping between them is normal practice rather than fussiness. The dead volume in the hub is also a real loss on small quantities.

Placebo arms in withdrawal designs are doing something unusual: they represent stopping, not never starting. Readers who carry an intuition from initiation trials will misread them, and the papers rarely spell out the difference.

The randomised withdrawal design answers one question well and is routinely quoted as answering a different one. It tells you what happens when treatment is removed from people who reached a given point on it. It does not tell you what happens to somebody who stops for their own reasons at their own moment.

Where people give more than one reason, most analyses take the first mentioned, which is an ordering effect rather than a weighting. Asking respondents to rank would cost nothing and would change several published distributions.

A small plea for the supplementary tables. The withdrawal literature puts its most useful numbers there — adherence, attendance, concomitant support — and the main text prints the curve everybody reproduces.

Monoisotopic and average mass differ by enough at peptide scale to matter, and the two are quoted interchangeably in supplier literature. For a forty-residue sequence the gap is a couple of daltons, which is exactly the size of the modification a buyer might be trying to exclude.

Your charge-state arithmetic is correct and I would set it out for readers as a rule they can apply. Take the observed mass-to-charge value, multiply by the charge, then subtract the charge multiplied by the mass of a proton. Two adjacent peaks in an envelope give you the charge and therefore the mass without any assumption at all.

The claim that a reproduced spectrum is worth more than any number in the document seems overstated. Most buyers cannot read a spectrum, and a printed image invites false confidence rather than scrutiny.