# Reference Cultures vs. Commercial QC Panels: Which Is Right

A reference culture is an authenticated living microorganism that a laboratory obtains from a recognized culture collection and maintains on its own bench, while a commercial QC panel is a ready-to-use set of organism preparations that a manufacturer grows, packages, quantitates, and ships for a defined assay. The rule of thumb is straightforward. Buy commercial panels for the high-frequency checks you run every day, and keep authenticated reference cultures for the checks you must defend, such as method verification, antimicrobial susceptibility testing (AST) quality control, and investigations of drifting results. Most laboratories that handle cultures need both, and the real decision is which job each tool carries.

Quality control (QC, the routine testing of known material to confirm that a method performs correctly) works differently in microbiology than anywhere else in the laboratory, because the control material is alive. A chemical standard in a sealed bottle behaves the same way for months. A bacterial strain divides, mutates, sheds plasmids (small self-replicating DNA elements that can be lost during storage and transfer), and dies if someone forgets it. That single fact drives the question of reference cultures vs. commercial QC panels, and it explains why laboratory quality management frameworks treat living control material differently from every other reagent on the shelf. Deciding on reference cultures or commercial QC panels for each job is an assignment exercise, not a search for a single winner.

This article defines each option on its own terms, compares them side by side on the criteria that decide real purchases, walks through four bench-level scenarios, and lists the mistakes laboratories make when they pick one and ignore the other.

## What Reference Cultures Are

A reference culture, often called a reference strain or control strain, is a microorganism whose identity has been verified by the institution that deposited it and whose passage history is documented from the original deposit. Reference strains are distributed by culture collections such as the American Type Culture Collection (ATCC), the National Collection of Type Cultures (NCTC), the Leibniz Institute DSMZ, and the Institut Pasteur collection, usually as lyophilized (freeze-dried) vials or frozen stocks. A special subset, the type strain, is the single named strain on which a species name is formally anchored in nomenclature. When a laboratory reports a name such as Pasteurella multocida or Salmonella enterica, the name ultimately traces back to material of this kind held in a long-term collection.

### What a reference strain controls for

Growth support and sterility of media. A strain with known, demanding growth requirements, inoculated onto each new lot of culture medium, confirms that the lot supports growth. A fastidious organism (one with complex nutritional needs, such as Haemophilus influenzae) catches lots that grow easy organisms but fail hard ones. Uninoculated lots held under normal incubation conditions serve as sterility checks.

Stains and reagents. A known Gram-positive organism and a known Gram-negative organism verify each new batch of Gram stain reagents, oxidase reagent, and catalase reagent before those reagents touch patient or field isolates.

Identification systems. Reference strains run through biochemical panels, MALDI-TOF workflows (matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, which identifies a microorganism from its ribosomal protein mass fingerprint), and sequencing pipelines confirm that the system returns the right name for a known input.

AST quality control. Standard strains such as Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 25923, and Pseudomonas aeruginosa ATCC 27853 carry published acceptable ranges for zone diameters in disk diffusion (antibiotic-impregnated disks on an inoculated agar plate, read as the diameter of the zone of inhibited growth) and for minimum inhibitory concentrations (MICs, the lowest antimicrobial concentration that visibly inhibits growth). When the control strain lands inside its range, the medium, the inoculum density, the incubation conditions, and the reading technique are all in control at once. These strains sit at the center of microbiology quality control worldwide.

Molecular assay controls. A characterized strain supplies extraction and amplification control material for [polymerase chain reaction](/knowledge/molecular-biology/polymerase-chain-reaction) (PCR) panels, and its nucleic acid behaves predictably across runs.

### How laboratories keep reference strains alive and honest

Standard practice separates the master stock (the original vial or a first-generation stock, stored lyophilized or frozen at around -80 C in glycerol broth) from working stocks (small aliquots derived from the master for daily use). Working stocks are limited to a small number of subcultures (transfers to fresh medium), commonly five or fewer, and are then discarded and replaced from the master. Each subculture receives a purity check and, where relevant, a spot identity check. The laboratory logs receipt date, lot number, passage number, storage location, and QC results for every stock.

Drift is the standing threat. Any living stock can change genetically over generations, and laboratories that maintain biological materials treat genetic monitoring as a formal element of quality control. Researchers who maintain inbred laboratory mouse colonies, for example, verify colony identity by mining genome-wide microsatellite markers (short repeated DNA motifs whose length varies between lineages) from double digest restriction-site associated [DNA sequencing](/blog/guides/dna-sequencing) (ddRAD), and they describe this genetic monitoring as one of the key elements of quality control and colony management [1]. Bacterial reference stocks deserve the same suspicion, because the underlying biology is identical.

### What "characterized" should mean

Modern reference strains increasingly carry whole-genome sequence data, and that data sets the benchmark any laboratory should demand from a supplier of biological control material. One 2025 dataset sequenced the complete genomes of five live attenuated bacterial vaccine strains used in South Korea ([Erysipelothrix rhusiopathiae](/knowledge/bacteria/livestock-bacteria/erysipelothrix-rhusiopathiae) NL-11, [Clostridium chauvoei](/knowledge/bacteria/livestock-bacteria/clostridium-chauvoei) B, two [Bordetella bronchiseptica](/knowledge/bacteria/pet-bacteria/bordetella-bronchiseptica) strains, and Mycoplasmoides gallisepticum str. F) with a hybrid workflow that combined Illumina short reads on a NovaSeq 6000 with Oxford Nanopore long reads on a MinION, producing gap-free, annotated genomes between 0.95 and 5.34 megabases [2]. A companion dataset covered eleven strains used in inactivated vaccines, including Pasteurella multocida, [Actinobacillus pleuropneumoniae](/knowledge/bacteria/livestock-bacteria/actinobacillus-pleuropneumoniae), Glaesserella parasuis, [Mannheimia haemolytica](/knowledge/bacteria/livestock-bacteria/mannheimia-haemolytica), Avibacterium paragallinarum, and three Escherichia coli strains, with assemblies more than 99 percent complete and less than 1 percent contaminated [3]. A name on a vial is a claim. Sequence-level identity is evidence.

## What Commercial QC Panels Are

A commercial QC panel is a set of organism preparations that a manufacturer produces under lot control, typically lyophilized or frozen, often quantitated to a target concentration, and supplied with a certificate of analysis (a lot-specific document listing the organism, its expected performance, and its expiry). The laboratory reconstitutes or thaws the material, runs it through the assay, and compares the result against the manufacturer's stated acceptance criteria.

### Common formats

- Lyophilized QC organisms for daily bench QC, supplied as discs, swabs, or vials for AST checks, media growth checks, and reagent verification, with no stock maintenance required.
- Quantitated suspensions standardized to a defined cell density, which removes the most error-prone manual step in microbiology QC, adjusting an inoculum to a 0.5 McFarland turbidity standard (the reference suspension used to standardize bacterial inocula).
- Positive and negative controls for PCR and sequencing panels, increasingly synthetic or inactivated so the material carries no infection risk and no viability.
- Calibration and extraction-control organisms for MALDI-TOF systems.
- External proficiency testing panels, coded specimens sent on a schedule by an outside provider and graded against a peer group. This is the panel concept applied across laboratories rather than within one.

### The panel concept at inter-laboratory scale

The clearest picture of what panel-based testing can and cannot show comes from a ten-year rabies program. From 2009 to 2019, 73 laboratories, 59 percent of them in Europe, took part in proficiency tests on animal specimens using four techniques: the fluorescent antibody test (FAT, a microscopy method that uses fluorescently labeled antibodies to detect viral antigen in brain tissue), the rabies tissue culture infection test (RTCIT, which inoculates cell cultures and examines them for viral infection), conventional RT-PCR (reverse transcription polymerase chain reaction, which copies viral RNA into DNA before amplification), and real-time RT-PCR. Conventional RT-PCR produced the highest agreement with the true panel results, 99.3 percent (95 percent confidence interval, the range that would contain the true value in 95 of 100 repeated samples, 99.0 to 99.6), followed by FAT at 99.1 percent, real-time RT-PCR at 98.7 percent, and RTCIT at 96.8 percent [4]. Because the panels were prepared from experimentally infected animals, the organizers could also estimate each technique's diagnostic sensitivity (how often it detected true positives) and specificity (how often it correctly cleared true negatives) directly, something a laboratory's own daily controls can never provide [4].

That last point deserves emphasis. A daily QC organism tells you that your assay worked today on your bench. A graded external panel tells you how your result compares with everyone else's and, when the panel's true composition is known, how your method performs in absolute terms. The two answers are not interchangeable.

### Where commercial panels stop

Commercial panels have three structural limits. First, the menu is fixed. A vendor sells controls for the assays it sells, so a laboratory validating an unusual method must look elsewhere. Second, identity rests on the manufacturer's chain of custody. The laboratory receives documentation, not independent proof. Third, inactivated or synthetic material cannot verify steps that depend on viable organisms. A medium growth check needs a living cell. Panels are excellent for what they cover and silent about everything else.

## Reference Cultures vs. Commercial QC Panels: Side-by-Side

The difference between reference cultures and commercial QC panels comes into focus when the two options sit on the same criteria.

| Criterion | Reference cultures | Commercial QC panels |
|--|--|--|
| What they are | Authenticated living strains the laboratory maintains itself | Manufacturer-prepared organism preparations, ready to run |
| Typical source | Culture collections such as ATCC, NCTC, and DSMZ | Assay and instrument manufacturers, proficiency testing providers |
| Identity assurance | Documented provenance, increasingly backed by whole-genome sequences | Manufacturer lot documentation and certificate of analysis |
| Traceability | Traceable to deposited strains, often type strains | Traceable to the manufacturer's seed and production records |
| Organism coverage | Any strain the laboratory can obtain and legally handle | Fixed menu tied to the vendor's assay list |
| Labor demand | High: subculturing, passage limits, purity checks, records | Low: reconstitute, run, record |
| Cost pattern | Low cost per organism, high hidden labor cost | Higher cost per test, labor shifted to the manufacturer |
| Documentation | Built and maintained by the laboratory | Supplied with each lot |
| Best suited to | Method verification, AST QC anchors, training, drift investigations | High-frequency daily QC, molecular controls, external assessment |
| Main failure modes | In-house drift, contamination, mislabeling, storage loss | Shipping damage, lot-to-lot variation, limited strain diversity |
| Turnaround after a QC failure | Fast if a working stock is on hand, slower if reviving a frozen stock | Fast, reconstitute a fresh unit |
| Typical user | Full-service clinical, veterinary, and biologics laboratories | High-volume laboratories and networks with limited bench time |

Three rows decide most purchases: identity assurance, organism coverage, and labor demand. The next section takes them one at a time.

## How to Choose: The Criteria That Decide Real Choices

### Identity assurance and traceability

Reference cultures win this criterion on paper, because their provenance is documented and increasingly genomic. The vaccine strain genome datasets illustrate the endpoint. Their stated purpose was to provide a reference for verifying, at the strain level, whether manufacturers keep using the same seed strains over time, and to support consistent mass production and reliable quality control of biological products [2]. Datasets built to that standard give a laboratory a fixed identity anchor that survives changes in staff, reagent lots, and instruments.

Commercial does not automatically mean authenticated. The clearest demonstration comes from brucellosis vaccine quality control. Investigators typed 63 commercial batches of Brucella abortus vaccine, 53 batches of the S19 strain from seven manufacturers and 10 batches of the RB51 strain from two manufacturers, all produced between 2006 and 2009, using multiple-locus variable-number tandem repeat analysis (MLVA16, a method that counts repeating DNA sequences at 16 genomic loci to fingerprint a strain). They also passed reference strains and selected commercial batches through ten serial in vitro passages. The vaccine strains proved genetically stable and highly homogeneous overall, but batches from one S19 manufacturer and batches from one RB51 manufacturer carried genotypes distinct from the reference vaccine strains, differing at a single locus, Bruce07, by one repeat unit [5]. A one-repeat difference is invisible to every biochemical and phenotypic test in a routine laboratory. Only comparison against the reference strain, at the genomic level, caught it.

Strain identity also underwrites the claim attached to a biological product. The Malaysian Society of Gastroenterology and Hepatology built its position statement on probiotics around strain-specific efficacy and safety, covering quality control in manufacturing, criteria for selecting products, and the evidence for specific strains, and issued twelve statements after reviewing the literature from 2011 to 2020 [6]. A probiotic whose contents do not match the named strain cannot deliver the evidence behind that strain. The same logic applies to every control organism. The value of the control is the identity of the strain in the vial.

### Organism coverage and flexibility

Reference cultures cover anything the laboratory can obtain and legally handle: fastidious organisms, rare serotypes (surface-antigen variants of a species), locally relevant pathogens, and resistance phenotypes that no off-the-shelf panel includes. A laboratory validating a new Campylobacter method or a Brucella identification pathway will almost always build its controls from authenticated strains, because no vendor sells a panel for that exact combination. Commercial panels cover what the vendor sells, which is exactly right for routine menus and exactly wrong for anything unusual.

### Labor, cost, and turnaround

Reference cultures demand bench time on a schedule: subcultures, purity plating, turbidity adjustment, passage logging, and recovery from the occasional storage failure. Commercial panels compress all of that into reconstitution and running, and they shift the labor bill to the manufacturer. The crossover depends on volume. High-frequency daily QC favors panels, because the labor cost of maintaining daily in-house stocks usually exceeds the panel price. Low-frequency verification, unusual organisms, and investigative work favor in-house reference stocks, because buying panels for occasional needs wastes money and still leaves coverage gaps. Turnaround matters too. After a QC failure blocks reporting, a lyophilized panel unit that reconstitutes in minutes returns the assay to control faster than reviving a frozen stock, unless the laboratory keeps a fresh working stock on hand, which is precisely what good reference culture discipline provides.

### What accreditation and standards expect

Laboratory quality management standards for medical laboratories, such as ISO 15189, expect documented method verification, lot control of reagents, and participation in external quality assessment (also called proficiency testing). Standards bodies such as the Clinical and Laboratory Standards Institute publish the acceptable QC ranges for AST control strains and the schedules for testing them, which typically begin with testing every day or every run and allow reduced frequency after a documented series of acceptable daily results (a common schedule moves a laboratory from daily to weekly AST QC after 20 to 30 consecutive in-control days). External proficiency testing then adds the layer that daily QC cannot provide. The ten-year rabies program showed why that layer matters, because panel-based grading exposed technique-specific weaknesses that no single laboratory could have seen from its own daily controls [4].

## When to Use Reference Cultures

Use a reference culture when the result must be defensible, when the organism is unusual, or when you need a fixed point to investigate a problem.

Verifying a new method, instrument, or reagent lot. Before a new MALDI-TOF, a new nucleic acid extraction system, or a new lot of culture medium enters service, run authenticated strains with known behavior, record the results, and release the method against that record.

Anchoring AST QC. The standard control strains and their published ranges are the reference point for every zone size and MIC the laboratory reports. This is the most heavily standardized QC requirement in [clinical microbiology](/knowledge/diagnostics/microbiology/clinical-microbiology-from-specimen-collection-to-pathogen-identification), and it is built on reference cultures.

Investigating discrepant or drifting results. When a zone size drifts or an identification flips, the reference strain is the fixed point that separates an organism problem from a reagent problem from a technique problem. You control every variable, which is exactly what an investigation needs.

Training and competency assessment. New staff run known organisms before unknown ones, and competency assessments use material with known answers.

Covering assays with no commercial panel. Rare, fastidious, or locally relevant organisms, and resistance mechanisms that panels do not include, end up as in-house controls built from authenticated strains.

Any claim that depends on strain identity. The probiotics position statement makes the point plainly: efficacy and safety are strain-specific properties, so a product's value collapses if the strain in the container is not the strain on the label [6]. Laboratories that make claims about their own materials, from medium performance to biologics seed stocks, carry the same burden.

## When Commercial QC Panels Are the Better Fit

Choose panels when frequency, standardization, or safety drives the decision.

High-frequency daily QC with limited staff. If the alternative is skipping QC on a busy day, the panel wins, because a skipped QC run is worse than an expensive one.

Molecular assay controls that need lot-to-lot consistency. Quantitated, lot-controlled control material behaves the same way across months, which is what trend analysis needs.

Laboratories without stock-keeping capacity. No freezer space, no maintenance expertise, no backup if a stock dies: panels remove the entire stock-keeping burden.

Reduced handling of hazardous organisms. Inactivated and synthetic controls remove exposure risk for agents such as Brucella species or rabies virus. Keep the boundary in mind: inactivated material cannot verify viability-dependent steps, so culture-based checks still need live organisms.

External proficiency testing. By definition, inter-laboratory comparison arrives as a panel, and accreditation expects it. A laboratory cannot substitute its own reference strains for this layer, because the value comes from the comparison, not the organism.

## Worked Scenarios

### Scenario 1: A small clinical microbiology laboratory sets up AST QC

A three-person laboratory reports disk diffusion and MIC results for routine aerobic organisms. The director buys lyophilized units of the three standard AST control strains for daily QC and keeps one authenticated stock of each strain as a reference backup. For the first 20 to 30 days, QC runs daily. All results fall in range, so the schedule moves to weekly under the published reduction scheme. When a new medium lot arrives, the control strains run on the new lot before the switch. Six months later, a zone size drifts out of range. The laboratory reruns a fresh panel unit, then reruns its own reference working stock, and the two disagree. The panel unit is in range and the in-house stock is not, which points to drift or contamination in the working stock. The laboratory discards the working stock, restarts from the master, and closes the event with a documented root cause. Each tool did a job the other could not.

### Scenario 2: A veterinary diagnostic laboratory adds a PCR panel

A veterinary laboratory adds a real-time PCR panel for a respiratory pathogen. The kit ships with inactivated positive control material matched to each lot. For daily runs, the laboratory uses the kit control, which is the right tool, because it is lot-matched and safe to handle. For initial verification, the laboratory obtains nucleic acid from a well-characterized strain of the target organism and confirms the assay detects it before the kit control enters routine use. The division of labor is clean. The reference material proved the assay works. The commercial control proves each kit lot still works. When the laboratory later joins an external proficiency round for the same pathogen, the graded result closes the loop by comparing the laboratory against its peers, the layer that neither daily controls nor in-house reference strains provide.

### Scenario 3: A biologics producer audits its seed strain system

A producer of bacterial biologics wants to demonstrate that its seed stocks have not changed across production runs. The audit has three layers. First, the producer sequences its master seed stocks completely and keeps the annotated genomes as a fixed reference, following the practice now published for both live [2] and inactivated [3] veterinary vaccine strains, so that any future batch can be compared against a stable genomic anchor. Second, the producer fingerprints each production batch against the seed reference with a locus-level method such as MLVA, which has been shown to discriminate single-repeat differences between vaccine batches and their reference strains and has been recommended for inclusion in official in vitro quality control tests [5]. Third, routine release testing continues to use commercial QC panels and standard control strains that anchor daily media and reagent checks. The reference cultures are the backbone. The panels are the daily plumbing. Neither substitutes for the other.

### Scenario 4: A laboratory chases a drifting zone size

A laboratory's Escherichia coli ATCC 25922 zone sizes for one antimicrobial have crept downward over eight weeks but remain in range. Nothing failed, so nothing triggered an investigation. The correct move is trend review, not waiting for failure. The laboratory plots the zone sizes on a control chart (a plot of QC results over time against fixed limits), sees the drift, and works through the fixed points: rerun the reference working stock from a fresh subculture, check the medium lot, check the inoculum turbidity, check the incubator temperature log. If the reference stock reproduces the drift, the problem sits in the reagents or the technique. If the reference stock is stable and only the panel units drift, the problem sits in the panel lot, and the certificate of analysis and vendor follow-up close it. The reference culture is the diagnostic instrument for the QC system itself.

## Common Mistakes and Limitations

Assuming commercial means authenticated. A certificate of analysis documents what the manufacturer tested, not an independent verification of identity. The brucellosis vaccine study showed that even licensed biological products from established manufacturers can carry genotypes that differ from their reference strains [5]. Buy panels for convenience, and verify identity for anything that anchors a claim.

Subculturing without limits. Every passage is a chance for drift, contamination, or a labeling error. Keep the master stock untouched, cap working stocks at a small number of passages, and log every transfer.

Confusing daily QC with proficiency testing. Daily controls check your assay on your bench today.

## Common Mistakes and Limitations, Continued

The draft's list stops mid-item, so this section completes it and adds the failure modes that surface most often when a laboratory quality management program is audited.

Confusing daily QC with proficiency testing. Daily controls check your assay on your bench today, and they only test the organisms you chose to run. That design carries a structural blind spot: it can never show how the method performs on material the laboratory did not select, and it can never show how the laboratory's answers compare with anyone else's. A graded external panel removes both blind spots at once. The ten-year rabies program could rank four techniques against each other, conventional RT-PCR at 99.3 percent agreement with true panel results against 96.8 percent for the tissue culture infection test, only because 73 laboratories ran the same coded specimens [4]. No amount of daily in-house QC would have produced that comparison. The fix is scheduling, not substitution: both layers belong on the calendar, and proficiency panels must be run exactly the way routine specimens are run, because a round handled with special care measures the laboratory's best day rather than its ordinary one.

Treating QC as pass or fail instead of a trend. An in-range result says nothing about direction. A zone size can sit inside its published limits while sliding steadily toward the edge, and eight weeks of technically passing results can end in a failure that was visible at week two. The fix is a control chart with a named reviewer and a defined review interval, covered in detail below. The principle is simple: microbiology quality control lives in the trend line, not in the single result.

Buying reference strains and never building the stock system. A lyophilized vial in a drawer is a purchase, not a program. Without rehydration, master and working stocks, passage limits, purity checks, and records, the vial is a single-use item that will be opened under pressure during an investigation, by someone who has never handled it, with no baseline showing what the strain did when it was healthy. The stock system is what converts a purchased strain into a laboratory asset, and the step-by-step guide below lays it out.

Substituting patient or field isolates for authenticated controls. Recent isolates are convenient, free, and biologically realistic, which makes them tempting as control material. They are also unanchored. Their identity rests on the laboratory's own unverified calls, their behavior has no documented baseline, and their properties can change in storage exactly the way any living stock's can. An isolate can serve as supplementary challenge material once its identity is independently confirmed, but it cannot replace a strain whose provenance and expected performance are documented.

Assuming one control strain covers every condition. Escherichia coli ATCC 25922 growing well on a new medium lot proves that the lot supports Escherichia coli. It says nothing about whether the lot supports Haemophilus influenzae, and fastidious organisms fail lots that satisfy easy ones. Map every control strain to the specific attribute it verifies, and make sure the demanding attributes have a demanding organism assigned to them.

Letting the schedule erode under workload. The first casualty of a busy week is the QC run that could technically be deferred. Every deferred run is an interval in which the assay produced results without evidence of control, and the deferral is rarely documented, so the out-of-control window cannot even be reconstructed later. If staffing makes the schedule unrealistic, the honest fixes are reduced-frequency schemes where published ones exist, or panels that cut the labor per run, not silent skipping.

Ignoring shipping and storage excursions. Lyophilized panels arrive through uncontrolled logistics, and frozen stocks sit in freezers that lose power. Both events are manageable, but only if they are detected. Temperature indicators on shipments, continuous logging on storage equipment, and a quarantine step after any excursion turn a silent failure into a managed one. Material that has experienced an unquantified excursion is suspect until it proves otherwise.

Recording failures without closing them. A QC failure entered in a log with no root cause, no corrective action, and no effectiveness check is a recurring event on a delay. Accreditation frameworks expect closure, and closure means a documented cause, a documented change, and a later look to confirm the change worked. The troubleshooting walkthrough below is built to produce exactly that closure.

Over-trusting inactivated and synthetic material. Inactivated and synthetic controls are safer, more stable, and often more consistent than live material, and they cannot verify a single step that depends on viability. Growth support, colony morphology, biochemical reactions driven by live metabolism, and any culture-based confirmation all require a living cell. A laboratory that moves entirely to non-living controls has quietly stopped verifying its culture steps.

Running expired material. Expiry is part of the certificate of analysis, not a suggestion printed on a label. Expired control material invalidates the check it is used for, and a result released on an expired control is a result released without control. First in, first out rotation and a reorder point tied to usage rate prevent this without heroics.

Concentrating know-how in one person. If one scientist holds the revival procedure in their head, the program has a single point of failure shaped like a resignation letter. Written procedures, cross-training, and a location map that lets a stranger find any stock are what make the system survive staff turnover.

Finally, the honest limits of the whole framework. Neither tool verifies the pre-analytical phase: collection, transport, and handling of specimens happen before any control organism is touched, and no QC panel reaches back that far. Neither tool substitutes for analyst competency, because a control run by an untrained hand is a controlled assay performed badly. And neither tool can rescue a method that is fundamentally unfit for its purpose; QC confirms that a method is in control, while verification establishes that it was worth controlling in the first place. The difference between reference cultures and commercial QC panels matters enormously inside this framework and not at all outside it.

## Building a Reference Culture Program from Zero: A Step-by-Step Guide

**Step 1: Inventory the assays and write the control requirements list.** Walk every bench and list each assay, each reagent batch check, and each medium lot check. For each entry, decide what a control must prove: growth support, reagent performance, identification accuracy, quantitation, or viability. The output is one page listing every check with the strain or panel that serves it. This list is the contract between the bench and the quality system, and every later step hangs off it.

**Step 2: Select the strains.** Where published acceptance criteria exist, use strains that carry them; the standard AST control strains are the model, because their zone and MIC ranges are published and their behavior is documented worldwide. Where no published range exists, choose strains with documented behavior and record the laboratory's own acceptance criteria during verification. Deliberately include at least one fastidious organism for medium checks, because easy organisms pass lots that fail hard ones. Keep two roles mentally separate: type strains anchor species names in nomenclature, while routine control strains anchor daily performance. Both are reference cultures, but they do different jobs, and a laboratory rarely needs type strains on the bench.

**Step 3: Order from a recognized collection and quarantine on arrival.** Order from collections such as ATCC, NCTC, DSMZ, or the Institut Pasteur collection. On receipt, record the collection number, the supplier lot, the date, and the condition of the packaging. Quarantine the material until identity and purity are confirmed, exactly as a laboratory would quarantine a new reagent lot.

**Step 4: Rehydrate and establish the master stock.** Follow the supplier's insert for rehydration medium and conditions, streak for isolation, confirm a single colony morphology, and run the laboratory's identification method on the recovered organism. Only then prepare the master stock: multiple aliquots, stored lyophilized or frozen at around -80 C in glycerol broth. Aliquot so that using one master aliquot never requires thawing the set. A master that is repeatedly handled stops being a master.

**Step 5: Derive working stocks under a passage ceiling.** Prepare small working aliquots from a single master aliquot, cap the lineage at five or fewer subcultures, then discard and restart from the master. Purity check every generation, and run a spot identity check wherever the downstream assay depends on identity. The passage ceiling is not bureaucracy; every transfer is an opportunity for drift, contamination, or a labeling error, and the ceiling bounds the exposure.

**Step 6: Build the stock record.** One record per strain, containing at minimum: collection number, source lot, receipt date, master stock locations and aliquot count, working stock lineage with passage numbers, QC results per generation, disposal dates, and operator initials. The record should let someone who has never seen the strain reconstruct its entire history in five minutes.

**Step 7: Schedule maintenance and monitoring.** Assign a named person, define intervals for purity checks and identity spot checks, monitor storage temperature continuously with an alarm, and put a revival drill on the calendar. Researchers who maintain inbred laboratory mouse colonies treat genetic monitoring as a formal, scheduled element of quality control rather than an ad hoc reaction to a problem [1], and a bacterial stock program deserves the same discipline: scheduled, documented, and acted on.

**Step 8: Plan for loss before it happens.** Keep a duplicate master set in separate storage, keep the collection catalog numbers on file so any strain can be reordered, and rehearse the revival drill periodically. The drill tests the whole chain, not just viability: it confirms that the label matches the organism, that the organism still behaves as documented, and that the person performing the revival can do it from the written procedure alone.

## Running Commercial QC Panels Well: Selection, Scheduling, and Lot Control

**Step 1: Map every assay to its panel.** Build a table with one row per assay and columns for control material, format, frequency, acceptance criteria, supplier, current lot, and expiry. Gaps in the map are as informative as entries: an assay with no assigned control is a finding, and the map makes the gap impossible to overlook.

**Step 2: Write acceptance criteria before the first run.** Pull the criteria from the certificate of analysis and from any published ranges, and write them into the standard operating procedure. A result compared against nothing is an observation, not a control. If the manufacturer states an expected result only qualitatively, record what the laboratory will accept before the material is ever run.

**Step 3: Set frequency by risk.** High-volume and high-consequence assays justify daily QC. Reduced schedules apply where published schemes allow them; the common AST reduction moves a laboratory from daily to weekly testing after a documented series of in-control days, on the order of 20 to 30. Beyond the schedule, add trigger events: every new lot, every new operator until competency is signed off, and every return from instrument service or repair.

**Step 4: Control the lots.** On arrival, verify the certificate against the order and the label. Run a new lot before the old lot is retired, or in parallel with it, so a lot problem is caught before it blocks reporting. Record the lot number with every QC result, because trends that look like drift in a mixed plot often resolve into clean lot-to-lot steps once the data are stratified.

**Step 5: Respect the insert.** Reconstituted and thawed materials have defined handling windows. Stretching those windows converts a control into a variable, and the resulting wobble lands in the trend chart where it will eventually cost an investigation.

**Step 6: Trend by lot.** Plot results with lot boundaries marked, or maintain separate plots per lot. Lot-to-lot shifts are invisible in a mixed plot and obvious in a stratified one.

**Step 7: Manage expiry actively.** Rotate stock first in, first out, quarantine units at expiry, and set the reorder point from the usage rate rather than from the moment the last box is opened.

**Step 8: Agree on the escalation path with the vendor in advance.** Know who to contact, what documentation the vendor will request, and at what point a lot investigation becomes a review of results reported while the lot was in service. Negotiating this during a failure is the most expensive possible time to do it.

## Trending Living Controls: Charts That Actually Get Reviewed

A control chart is a plot of QC results over time against fixed limits, and for living controls it is the single highest-value document in the QC file. Living material adds biological variability on top of analytical variability, which widens the limits and makes any single result a weak signal. Drift, the thing that actually hurts laboratories, shows up only across weeks of results.

Build the chart with these elements. Plot every result: zone diameters, MIC values, growth ratings, or quantitated molecular values, whichever the assay produces. Draw the published or laboratory-defined limits as fixed lines. Annotate events directly on the chart: medium lot changes, panel lot changes, operator changes, instrument service visits, incubator incidents, and reagent batch changes. An unannotated chart forces every future investigator to reverse-engineer the timeline from memory.

Give the chart a named reviewer and a defined review interval. A weekly review by a senior technologist with authority to hold reporting, plus a monthly summary to the quality meeting, is a workable pattern for most laboratories. The reviewer is looking for four signals: a run of results sitting on one side of the center of the range, a steady drift toward a limit, widening scatter around the center, and a step change that coincides with an annotated event.

Interpret the measurement scale before interpreting the signal. Zone diameters are continuous measurements, so small drift is visible early. MIC results move in dilution steps, so a single step in either direction is ordinary variation, while repeated steps in one direction across consecutive runs are a real signal. Treating a one-step MIC shift as a failure manufactures false alarms; treating a repeated one-directional march as noise misses real drift.

The purpose of the chart is to trigger investigation while everything still technically passes. This is precisely the situation in the draft's fourth scenario, where zone sizes crept downward for eight weeks without a single failure. The chart is what converts "nothing failed" into "something is changing," and it does so early, when the investigation is still cheap.

## Troubleshooting a QC Failure: A Systematic Walkthrough

**Step 0: Contain.** If the failure is unambiguous, stop reporting for the affected assay before doing anything else. Containment first, diagnosis second.

**Step 1: Confirm the failure is real.** Repeat with a fresh unit from the same lot. Along the way, check the obvious: expiry date, storage conditions of the unit, the reading itself, and any transcription between the bench and the record. A meaningful fraction of QC failures die at this step.

**Step 2: Bracket the problem with the reference stock.** Run the in-house reference working stock through the same assay. The four possible outcomes sort the world cleanly:

| Outcome | Interpretation | Next move |
|-|-|-|
| Panel unit fails, reference stock in range | Suspect the panel unit or its lot | Check certificate and shipping history, contact the vendor, stratify the trend chart by lot |
| Both fail | Problem sits in the assay environment | Isolate variables one at a time |
| Reference stock fails, panel unit in range | Suspect the in-house stock | Check purity and identity, restart from the master |
| Both in range on repeat | Transient event | Document it, watch the trend closely, look for what changed that day |

**Step 3: Isolate one variable at a time within the assay environment.** The usual suspects are the medium lot, the inoculum preparation and its turbidity standard, the incubation conditions, and the reading technique. Change one variable per run and never several, because stacked changes make the next failure uninterpretable. A second reader is the cheapest variable test available for any endpoint that involves human judgment.

**Step 4: Interrogate the organism itself.** Mixed colony morphology signals contamination. An unexpected identification signal suggests mislabeling or drift somewhere in the lineage. If the working stock is suspect, restart from the master. If the master is suspect, reorder from the collection; this is the moment the filed catalog numbers earn their keep.

**Step 5: Close the loop.** Record the root cause, the corrective action, and the review of any results reported during the out-of-control window. If patient or field results were released in that window, decide explicitly on retesting or amended reports and document the decision. An investigation without closure is a failure on a timer.

**Mini-example A: an oxidase check fails on a fresh reagent batch.** The known oxidase-positive control strain gives a weak reaction and the negative control behaves normally. Three variables are in play: the age of the prepared reagent, the passage history of the control strain, and the timing of the read. Reprepare the reagent fresh and rerun the reference strain from a fresh subculture. If the fresh combination is strongly positive, the old reagent had degraded and the fix is a shorter reagent life in the SOP. If the fresh combination is still weak, the working stock's expression has changed, and the restart point is the master stock. One variable at a time resolves it in two runs.

**Mini-example B: two changes land in one week.** A new incubator is commissioned and a new medium lot is released in the same week, after which the Staphylococcus aureus ATCC 25923 zone sizes shrink. Two variables changed at once, so no single comparison can assign the effect. The fix is to unstack them deliberately: run the reference strain on the old medium lot in the new incubator, and on the new lot under the previous conditions, and let the pattern of results assign the cause. The larger lesson is preventive: schedule commissioning and lot changes apart, because a confounded change is a drift investigation waiting to happen.

## Storage, Recovery, and Rescue of Reference Stocks

Master stock discipline starts with aliquoting. Freeze or lyophilize the master in many small aliquots so that any single use consumes one aliquot and touches nothing else. Every thaw-and-refreeze cycle is a stress event and an opportunity for contamination, so the design goal is that a master aliquot is opened once and then gone.

Labeling is a safety system, not a courtesy. Use labels that survive frost and condensation, and put the collection number, the passage generation, and the date on every vial. Color coding separates master from working stocks at a glance, and a location map, a simple document saying which box, rack, and position holds each strain, means any trained person can find any stock without asking anyone.

Storage monitoring is continuous, not periodic. Log temperature continuously, alarm on excursion, and connect the alarm to something a human will actually notice at 3 a.m., whether that is a dialer, a monitored system, or a documented after-hours response chain. Backup power for storage freezers is part of the same design.

When the freezer failure alarm fires, the sequence matters. Keep the door closed, because a closed freezer holds its temperature far longer than an opened one. Assess the duration and the peak temperature from the logging record. Move the highest-value material, which means master stocks, to backup storage first. After conditions are restored, purity check every revived stock before it returns to service, document the event and any stocks lost, and reorder from the collection any strain whose masters were all lost. The reordering step is why the collection numbers live in the stock records rather than in someone's memory.

The revival drill closes the loop. On a defined interval written into the SOP, revive a master aliquot, confirm growth and purity, confirm identity by the laboratory's method, and run the strain through its key performance check, for example the published AST range. If the drill passes, the master set remains valid. If it fails, the laboratory has discovered the loss during a scheduled drill instead of during an investigation, which is the entire point. Keep a lyophilized backup layer for the strains the laboratory cannot afford to lose, because lyophilized material tolerates storage accidents that frozen material does not.

## Fastidious, Hazardous, and Unusual Organisms: Special Handling

Fastidious organisms deserve a deliberate place in the control program. Keep at least one strain with demanding, defined growth requirements in the medium-check panel, because the whole value of a growth check is that it fails bad lots, and demanding strains fail lots that easy strains bless. Expect a shorter working stock life from fastidious lines, and write the growth requirements into the SOP so that the check fails loudly when the medium is wrong rather than quietly when the medium is merely mediocre.

Hazardous organisms call for a split decision. Use inactivated or synthetic material for molecular and serologic assay controls, where it removes exposure risk without losing control value. Reserve live authenticated strains for viability-dependent checks, and handle them only under the containment the organism demands. The brucellosis vaccine study explains why the live layer still matters: the genotype difference between certain commercial batches and their reference strains, a single repeat unit at locus Bruce07, was invisible to every routine phenotypic test and emerged only when batches were typed against the reference strains [5]. Live verification under proper containment is not nostalgia; it is the only layer that sees certain failures.

Unusual organisms are the natural territory of reference cultures. Locally relevant serotypes, rare resistance phenotypes, and organisms specific to a veterinary or biologics context rarely have off-the-shelf panels, so the laboratory builds controls from authenticated strains. When no published acceptance range exists, establish the laboratory's own criteria during verification, justify them in the verification record, and label them as laboratory-defined. A documented laboratory-defined range is defensible; an undocumented expectation is not.

## Verifying Identity: From Certificate to Genome

![Comparison of certificate, phenotypic confirmation, locus-level fingerprinting and whole-genome comparison by what each proves and when it is enough.](/article-images/illustrations/reference-cultures-vs-commercial-qc-panels-identity-layers.webp)

Think of identity assurance as four layers, each answering a stronger question than the last.

**Layer 1: the certificate and the collection number.** This documents the supplier's chain of custody and what the supplier tested. It is the entry ticket, not the proof.

**Layer 2: phenotypic confirmation on receipt.** The receiving laboratory streaks for purity, confirms a single morphology, and runs its own identification method. This answers a different question from layer 1: does the organism in this vial behave as labeled on this bench, with this laboratory's methods.

**Layer 3: locus-level fingerprinting.** Where a validated typing scheme exists for the organism, fingerprint the received strain and any in-house lineage against the reference. The brucellosis work is the demonstration case: MLVA16, which counts repeating DNA sequences at 16 genomic loci, discriminated commercial vaccine batches from their reference strains at a single locus, and the study's authors recommended the approach for inclusion in official in vitro quality control tests [5]. This layer is cheap relative to sequencing and catches differences that no biochemical panel can see.

**Layer 4: whole-genome comparison.** This is the modern benchmark, and the published vaccine strain datasets show the standard to demand: gap-free, annotated genomes produced by hybrid workflows that combined Illumina short reads on a NovaSeq 6000 with Oxford Nanopore long reads on a MinION, spanning 0.95 to 5.34 megabases for five live attenuated veterinary vaccine strains [2]. The companion dataset for eleven inactivated vaccine strains, including Pasteurella multocida, Actinobacillus pleuropneumoniae, Glaesserella parasuis, Mannheimia haemolytica, Avibacterium paragallinarum, and three Escherichia coli strains, applied explicit assembly quality gates of more than 99 percent completeness and less than 1 percent contamination [3]. A laboratory that sequences its master stocks once, to a comparable standard, creates a permanent identity anchor; every future working stock can be compared against that anchor for the life of the program.

Match the layer to the stakes. Receiving a new routine control strain justifies layers 1 and 2. Anything that anchors a claim, a biologics seed stock, a strain whose name carries the evidence behind a product, needs layers 3 or 4. The probiotics position statement makes the underlying rule explicit: efficacy and safety are strain-specific properties, so a claim travels

## Sources

1. [Comparison of ddRAD derived genome-wide SSR markers in outbred and inbred Swiss albino mice.](https://pubmed.ncbi.nlm.nih.gov/40350065/)
2. [High-quality genome assembly and annotation of live animal vaccine bacteria strains in South Korea.](https://pubmed.ncbi.nlm.nih.gov/40898044/)
3. [High-quality genome assembly and annotation of inactivated animal vaccine bacteria strains in South Korea.](https://pubmed.ncbi.nlm.nih.gov/42576204/)
4. [Multi-annual performance evaluation of laboratories in post-mortem diagnosis of animal rabies: Which techniques lead to the most reliable results in practice?](https://pubmed.ncbi.nlm.nih.gov/33544702/)
5. [Genetic stability of Brucella abortus S19 and RB51 vaccine strains by multiple locus variable number tandem repeat analysis (MLVA16).](https://pubmed.ncbi.nlm.nih.gov/23933375/)
6. [Use of probiotics in clinical practice with special reference to diarrheal diseases: A position statement of the Malaysian Society of Gastroenterology and Hepatology.](https://pubmed.ncbi.nlm.nih.gov/33490608/)