Criteria used to identify bacteria

Classification and identification are two different operations, and confusing them is the commonest mistake in this part of microbiology. Classification is the building of the scheme itself. Taxonomists characterise strains, arrange them into a hierarchy of ranks — species, genus, family, order, class, phylum, kingdom and domain — and give each accepted taxon a valid name under the binomial system, in which the genus is capitalised, the species is written in lower case, and the whole name is italicised. Modern bacterial taxonomy is described as polyphasic: the same organism is judged on phenotypic characters (morphology, staining, biochemistry, antigens), genotypic characters (DNA base composition, DNA–DNA hybridisation, gene sequences) and phylogenetic position, so that a new species is never proposed on a single test. Older schools built dichotomous keys, in which each test splits the unknown into two branches, and numerical taxonomy, in which many equal-weight characters are compared and strains are clustered by percentage similarity. Reference strains, called type strains, are deposited in public culture collections such as the American Type Culture Collection or the National Collection of Type Cultures, and their descriptions are summarised in reference works such as Bergey’s Manual, precisely so that a laboratory in one country and a laboratory in another are naming the same thing. Identification is the opposite direction of travel: an unknown isolate is placed into a taxon that already exists. In routine work it is a comparison rather than a discovery, because the laboratory performs a standard set of tests and matches the pattern against a table of known profiles. The clinical consequence is that a report may legitimately stop at different levels. “Escherichia coli” is a species-level identification; “coliform” or “Enterobacteriaceae” is a classification-level statement that is often all the prescriber needs in order to choose an empirical drug, whereas a species-level name matters when it changes therapy, as with Salmonella Typhi or a carbapenem-resistant Klebsiella. A nurse reading a report should therefore ask what level of certainty the name represents, because a genus-level answer often means that the laboratory could not separate two species with the methods available, not that the identification failed.

What a sequence cannot tell you. The short answer is that molecular methods answer a different question from the one the clinician most often asks. A polymerase chain reaction or a gene sequence detects nucleic acid. It tells you that DNA from an organism is present; it does not tell you whether that organism is alive, whether it is present in a number that matters, or which antimicrobial will kill it. Dead bacteria release DNA that persists for days, so a positive signal after successful treatment is common and cannot by itself prove failure.

Why susceptibility needs growth. Susceptibility testing, by contrast, requires a living isolate: disc diffusion measures the zone of inhibited growth, and broth dilution measures the minimum inhibitory concentration that stops visible growth, and neither can be performed on a sequence. Only a few resistance mechanisms are predictable from a single gene — the mecA gene of methicillin-resistant Staphylococcus aureus and the blaKPC carbapenemase gene are good examples — while many resistances arise from several genes acting together, from efflux pumps or from loss of porins, so genotype and phenotype do not always agree.

Mixed specimens and enrichment. Culture also separates a mixture: a sputum or wound sample may contain three or four species, and a molecular test performed directly on it gives a combined signal, whereas the plate physically isolates each one. It concentrates a minority organism: a stool culture for Salmonella uses selective enrichment because the pathogen may be outnumbered a thousandfold by normal flora, and the classical threshold of 10⁵ colony-forming units per millilitre of urine remains a useful guide to separate significant bacteriuria from contamination, although lower counts can be significant in a symptomatic patient.

Where culture is still the standard. Culture produces the pure colony that MALDI-TOF and whole-genome sequencing then identify, and it supplies the live isolate that a reference laboratory needs for outbreak typing, for vaccine work and for the frozen biobank that future investigations will use. It remains the reference standard for Mycobacterium tuberculosis, and in many laboratories it is still the only affordable option, since molecular platforms need reagents, stable electricity and a cold chain. Finally, a molecular panel detects only what its primers are designed to detect, and inhibitors in blood or pus can suppress the reaction and produce a false negative. The practical conclusion is not that one method replaces the other but that they are complementary: molecular tests shorten the time to a name, while culture preserves the living organism on which treatment decisions still rest.

Everything in this topic is beyond the supplied notes, which stop at the phrase “specialized molecular biology technique”. The polymerase chain reaction copies a chosen stretch of DNA. The reaction mixture contains the sample DNA, two short primers that flank the target, the four nucleotide building blocks, a thermostable DNA polymerase (Taq, originally obtained from Thermus aquaticus) and a buffer. The mixture is placed in a thermal cycler and taken through repeated cycles: about 95 °C separates the two DNA strands, a step called denaturation; about 50–65 °C lets the primers bind their complementary sequence, called annealing; and about 72 °C lets the polymerase extend the primers, called extension. Each cycle doubles the target, so 30–40 cycles turn a few copies into millions, and it is that exponential increase which lets the method detect an organism present in very small numbers. In real-time PCR the accumulating product is measured by fluorescence, and the cycle at which the signal crosses the threshold — the Ct value — falls as the starting amount of target rises, which makes the test roughly quantitative. For identification the usual target is the gene for 16S ribosomal RNA, roughly 1 500 base pairs long. It contains conserved regions, nearly identical across most bacteria, which allow universal primers to amplify it from almost any species, and nine hypervariable regions lying between them, whose differences carry the species signal. The product is sequenced and compared with curated databases; a similarity of at least about 98.7 % to a type strain is commonly accepted as species-level agreement, while 95–98.7 % suggests only a genus-level placement. Those figures are conventions rather than laws: the older 97 % threshold proposed in 1994 was replaced as more sequences accumulated, which is itself a lesson in how identification depends on the reference database. The same 16S approach, applied by Carl Woese in the 1970s, revealed that the archaea form a separate domain of life, and in clinical work it identifies organisms that culture has failed to grow, such as the cause of culture-negative endocarditis. MALDI-TOF mass spectrometry works differently and much faster. A colony is mixed with a matrix, dried on a target plate and struck by a laser; the matrix absorbs the energy and ionises predominantly ribosomal proteins, which travel through a vacuum tube and are separated by the time they take to reach the detector. The resulting spectrum of mass-to-charge peaks is a reproducible fingerprint, and software compares it with a library of reference spectra and returns a score: about 2.0 or above is accepted as species-level, roughly 1.7–2.0 as genus-level. Identification of a colony then takes minutes instead of a day, which is why a nurse may learn that a blood culture has grown Escherichia coli within hours of the bottle flagging positive. Neither method replaces susceptibility testing: both name the organism, but both still begin from a grown colony.