How Research Compounds Are Discovered and Named

A 1965 paper in Nature carries the title “Inhibition of cell division in Escherichia coli by electrolysis products from a platinum electrode”. The agent that stopped the bacteria dividing was not a designed molecule. It was a product of electrolysis at a platinum surface, characterised only after the biological effect had already been seen.

Effect first, structure second. That is one of several ways research compounds are discovered, and the route a molecule took leaves permanent traces in its paperwork: in the shape of its skeleton, in the alphanumeric code it carried before anyone gave it a name, and in the registry number it keeps for the rest of its existence. This article is the map of that territory – the routes, the definitions, and the identifiers – with the individual cases treated in depth elsewhere in this category.

How research compounds are discovered, and how they are named

Four broad routes put a new small molecule into the scientific record. They are not mutually exclusive, and most real programmes mix them, but they answer different questions and they fail in different ways.

Route	What comes first	What the route is good at	Where it is weak
Accident and observation	An unexpected biological or physical effect	Finding mechanisms nobody was looking for	Cannot be planned, scheduled or scaled
Phenotypic screening	A measurable change in a cell or organism	Delivering activity in a real biological system	The molecular target may stay unknown for years
Target-based design	A defined protein or binding site	Rapid, interpretable structure-activity work	Activity at the target need not translate onward
Natural product isolation	An extract with an activity	Chemical novelty that no chemist would draw	Supply, complexity and difficult total synthesis

The balance between these is not a matter of opinion. Swinney and Anthony catalogued the discovery strategy behind every new molecular entity and biologic approved by the US Food and Drug Administration between 1999 and 2008. Of 259 approved agents, 75 were first-in-class with a new molecular mechanism of action; 50 of those were small molecules and 25 were biologics. Among the first-in-class small molecules, phenotypic screening accounted for 28 and target-based approaches for 17 – in a period when target-based work dominated the field’s attention and budget.

Natural products remain a structural source rather than a historical footnote. Newman and Cragg, reviewing approvals from January 1981 to September 2019, report that for cancer drugs approved between 1946 and 1980, 40 of 75 small molecules – 53.3 % – were natural products or natural-product derivatives. Their central conclusion is that the use of natural product structures, or synthetic variations on them, is “still alive and well”.

Effect before structure: what the cisplatin record shows

The 1965 Nature report by Rosenberg, Van Camp and Krigas is a clean example of the first route. The activity was observed; the responsible species was then isolated and identified. The compound that emerged is a square-planar platinum(II) coordination complex. PubChem lists it as CID 5702198 with molecular formula Cl₂H₆N₂Pt and a molecular weight of 300.05.

That entry carries a second lesson, and it is a structural one. The trans isomer has its own PubChem record, CID 441203, with the same atom count and a different three-dimensional arrangement of the two chlorides and two ammine ligands around the metal. Same formula, same elemental analysis, different compound. Geometry, not composition, is doing the work – which is exactly why a molecular formula is never an identity statement on its own.

Why an active metabolite becomes a compound in its own right

A fourth pathway rarely gets its own heading: the compound that an organism makes from another compound. PubChem describes fexofenadine as the major pharmacologically active metabolite of terfenadine. The two are catalogued separately, and comparing their entries shows precisely what changed.

	Terfenadine	Fexofenadine
PubChem CID	5405	3348
Molecular formula	C₃₂H₄₁NO₂	C₃₂H₃₉NO₄
Molecular weight	471.7	501.7
CAS Registry Number	50679-08-8	83799-24-0
Terminal group on the aryl ring	tert-butyl	2-methylpropanoic acid

The carbon count is identical. Two hydrogens are gone and two oxygens have appeared – the arithmetic signature of one methyl group oxidised to a carboxylic acid. Everything else in the molecule, the diphenylmethyl-piperidine and the butanol chain, is untouched. Metabolism performed a single, selective oxidation that a synthetic chemist would have found awkward to do on that position with that selectivity.

For a laboratory, this class of pair is unusually instructive. Two compounds share a skeleton, differ by one functional group, and differ sharply in polarity and acidity. They are the natural worked example for retention behaviour, for ionisation in mass spectrometry, and for why a reference standard has to be the specific species, not merely the right skeleton.

Scaffold and substituent: what actually changes down an analogue series

A scaffold is the connected core framework of a molecule – the ring systems and the linkers that hold them together. A substituent is what hangs off that framework. The distinction is not cosmetic. Changing a substituent usually shifts properties along a continuum: lipophilicity, acidity, hydrogen-bonding capacity, metabolic vulnerability. Changing a scaffold usually starts a different series altogether.

The terfenadine-fexofenadine pair is a substituent change. The scaffold is identical, and so the two compounds behave similarly in most separations and differently in the few that resolve a carboxylic acid from an alkyl group. This is why analogue series are dangerous to characterise casually: compounds designed to be similar are similar, including in retention time and in ultraviolet absorbance, and the analytical burden of telling them apart falls on methods that respond to the difference rather than to the shared core.

The same logic explains a recurring shape in medicinal chemistry papers: a fixed core, a table of substituents, and a column of measured values. What is being explored is one axis at a time. Papers that change the scaffold and the substituents simultaneously produce results that are difficult to attribute to anything.

What a CAS Registry Number encodes, and what it does not

A CAS Registry Number is an identifier assigned by the Chemical Abstracts Service. Its documentation describes the format precisely: up to ten digits in three hyphen-separated groups, with two to seven digits in the first group, two digits in the second, and a single check digit in the third.

The check digit is arithmetic, not chemistry. Read the digits before the final hyphen from right to left, multiply them by 1, 2, 3 and so on, sum the products, and take the remainder on division by ten. For terfenadine, 50679-08-8: the digits 8, 0, 9, 7, 6, 0, 5 weighted 1 to 7 give 8 + 0 + 27 + 28 + 30 + 0 + 35 = 128, and 128 modulo 10 is 8, which is the published check digit. The same calculation reproduces the check digits of fexofenadine (83799-24-0) and of the example 107-07-3 used in the CAS documentation.

What a CAS number does not encode is anything about the substance. It is not a structure, not a class, not a hierarchy. Two adjacent registry numbers have nothing chemically in common. Conversely, chemically close things receive different numbers whenever the registered substance differs: a free base and its hydrochloride, an enantiomer and its racemate, a hydrate and the anhydrous form. That is a feature. It means a CAS number identifies the material in the bottle rather than the sketch on the whiteboard – provided the number on the label is the one that matches the material, which is a question for the certificate of analysis rather than for the catalogue.

The check digit is worth ten seconds of anyone’s time. It will not catch a number that is valid but wrong for the substance. It will catch a transposed pair of digits, which is the commonest way a CAS number gets corrupted between a paper, a purchase order and a label.

From laboratory code to INN: what a name tells you

Before a compound has a name it has a code: a few letters, often the originating laboratory, followed by a number. Codes are internal bookkeeping. They are useful precisely because they carry no claim – a code says “this material”, not “this class of agent”.

Where a name does exist, it is frequently an International Nonproprietary Name selected under the World Health Organization’s INN programme, and it is built to be read. The WHO’s stem document states that “stems define the pharmacologically related group to which the INN belongs”, and that “whenever possible, an INN should include the ‘common stem’ expressing the pharmacologically-related group to which the substance belongs”. Names that would suggest an anatomical, physiological, pathological or therapeutic claim are avoided. The general principles add that INNs “should be distinctive in sound and spelling” and “should not be liable to confusion with names in common use”.

A handful of stems, with the WHO’s own definitions, shows how much information sits in the last syllables of a name:

Stem	WHO definition
-olol	β-adrenoreceptor antagonists
-tinib	tyrosine kinase inhibitors
-afil	inhibitors of phosphodiesterase PDE5 with vasodilator action
-prazole	antiulcer, benzimidazole derivatives
-oxacin	antibacterials, nalidixic acid derivatives
-vastatin	antihyperlipidaemic substances, HMG CoA reductase inhibitors
-mab	monoclonal antibodies

The part in front of the stem is deliberately arbitrary – a distinctive prefix whose job is to be unmistakable, not descriptive. So a name ending in -tinib tells you the pharmacological family and nothing about the chemistry; the ring system, the substituents and the salt form are all invisible at that resolution.

Most research compounds never reach this stage. A compound that was never developed for human use has no INN, and often no trivial name that more than a few hundred people would recognise. For those substances the CAS Registry Number, the molecular formula and a structure-derived hash such as an InChIKey are the only stable handles, which is why catalogue entries and certificates for laboratory reference materials lean on them so heavily.

Which threads this category develops separately

This page is a hub, and it deliberately stops short in several places. The individual discovery histories – what the primary literature actually records, as opposed to the version that circulates – are treated one compound at a time, because each turns on the detail of a specific paper. Synthetic routes are handled separately, since the question of how a skeleton is assembled is orthogonal to the question of how it was found. Structural families get their own treatment, because a scaffold discussion only becomes concrete once there is a real series with real substituent changes to point at. And the analytical side – how identity and purity are actually demonstrated for any of these substances – belongs to the analytical methods category, not here.

Nothing on this page is a use recommendation. Every substance discussed is treated as an object of study and an analytical reference material; the terms governing that are set out in the site rules.

Frequently asked questions

Was cisplatin discovered by accident?

The primary record supports the sequence, not the folklore. The 1965 Nature paper by Rosenberg, Van Camp and Krigas reports inhibition of cell division in Escherichia coli by electrolysis products from a platinum electrode. The biological effect was observed first and the responsible platinum species was identified afterwards, which is the defining pattern of discovery by observation rather than by design.

How are CAS Registry Numbers assigned?

They are assigned by the Chemical Abstracts Service as identifiers, not as descriptions. The published format is up to ten digits in three hyphen-separated groups: two to seven digits, then two digits, then a single check digit. The number encodes no structural or chemical information, so two consecutive registry numbers may belong to entirely unrelated substances.

How do you check if a CAS number is valid?

Compute the check digit. Take the digits before the last hyphen, read them right to left, multiply by 1, 2, 3 and so on, add the products and divide by ten; the remainder must equal the final digit. For 50679-08-8 the weighted sum is 128, giving 8. The test catches transposed digits, but not a valid number attached to the wrong substance.

What does the suffix -tinib mean in a drug name?

It is an INN stem, and the World Health Organization defines it as denoting tyrosine kinase inhibitors. Stems mark the pharmacologically related group a name belongs to. They say nothing about the ring system, the substituents or the salt form, so two compounds sharing a stem can be chemically very different.

What is the difference between a development code and an INN?

A development code is internal laboratory bookkeeping – usually letters identifying the originating group followed by a number – and it makes no claim about what the compound does. An International Nonproprietary Name is selected through the WHO programme and is constructed from a distinctive prefix plus a common stem that signals the pharmacological group. Most research compounds never receive an INN.

Why is an active metabolite sometimes developed as a separate compound?

Because the metabolite is a different molecule with different properties. Fexofenadine, which PubChem describes as the major pharmacologically active metabolite of terfenadine, differs from its parent by a single oxidation: the same 32 carbons, two fewer hydrogens, two more oxygens, a tert-butyl group converted to a carboxylic acid. That one change alters polarity, acidity and chromatographic behaviour.

What is the difference between a scaffold and a substituent?

A scaffold is the connected core framework – the rings and the linkers holding them together. A substituent is a group attached to that framework. Changing a substituent moves properties along a continuum within one series; changing the scaffold generally begins a new series. Analogue tables in medicinal chemistry papers hold the scaffold fixed and vary substituents for exactly this reason.

Do cisplatin and transplatin have the same molecular formula?

They have the same atom composition and different geometry. PubChem catalogues them as separate records, CID 5702198 and CID 441203, differing in the spatial arrangement of two chloride and two ammine ligands around platinum. This is the clearest demonstration that a molecular formula is an incomplete identity statement for any compound where stereochemistry or geometry exists.

References

Research use only. Nonsensia Lab supplies analytical reference standards for laboratory and research applications. This article is published for scientific and educational purposes. It is not medical advice, it does not describe any use in humans, and nothing in it should be read as a recommendation to administer any substance to a person or animal.

Filed under: Chemistry & Discovery

Nonsensia Lab supplies the compounds discussed in this guide as analytical reference standards for laboratory and research use.

Where Research Compounds Come From: Discovery, Synthesis and Structure