Chapter 2

Chapter 2: Specimen Types, Storage Conditions and Quality Metrics — the WIA-BIO-019 Taxonomy

A biobank can only promise quality for material it can name precisely. This chapter walks through the fourteen specimen types, five storage conditions and type-specific quality metrics that WIA-BIO-019 defines in §2, §4 and §6.3, and explains the laboratory reasoning behind each threshold.

The Three Axes of the Taxonomy

Chapter 1 described a biobank in general terms: donors give biospecimens, the bank processes them into aliquots, stores them under controlled conditions and later distributes them to researchers. To turn that picture into something a laboratory information system can enforce, WIA-BIO-019 classifies every specimen along three independent axes. The first axis is what the material is, expressed by the SampleType enumeration. The second is how cold it is kept, expressed by StorageCondition. The third is how its fitness for use is measured, expressed by a family of quality interfaces (DNAQuality, RNAQuality, ProteinQuality, CellQuality and BloodQuality).

Keeping the axes separate matters. A plasma aliquot and a DNA extract can share a freezer shelf, but they are judged by completely different metrics; conversely, two DNA extracts may sit at different temperatures depending on how long they are expected to be kept. A data model that fuses type and temperature into one label ("frozen plasma", "fridge DNA") cannot represent a transfer from one freezer to another without also pretending that the material changed. WIA-BIO-019 avoids this by recording the three facts independently and letting policy rules connect them.

The reference TypeScript types define fourteen SampleType values: whole_blood, plasma, serum, buffy_coat, dna, rna, tissue_fresh_frozen, tissue_ffpe, pbmc, cell_line, urine, saliva, stool and the catch-all other. They also define five StorageCondition values: liquid_nitrogen, ultra_low_freezer, standard_freezer, refrigerated and ambient. The table below is the chapter's master reference. It condenses the scientific content of spec §2 and §6.3 and Appendix B into one row per enumeration value. The sections that follow explain where each number comes from and where it should be read with caution.

Two remarks before the details. First, "storage condition" in this table is the enumeration value a registration record would carry; it says nothing about how long the material may stay there. Second, every threshold in the table comes from the standard's own text. Where the spec's number is a convention rather than a law of nature, the discussion below says so.

2.1 Blood Products

Blood is the workhorse of most population and clinical biobanks because a single venipuncture yields several derivative types. WIA-BIO-019 models four of them as separate SampleType values. They are separate because they are processed differently, stored differently and fail in different ways.

Whole blood — WHOLE_BLOOD

§2.1.1 specifies collection into EDTA or citrate tubes, a typical volume of 2–10 mL, storage at −80 °C and a quality metric of "hemolysis index < 50". The anticoagulant choice is not cosmetic. EDTA chelates the calcium ions that the coagulation cascade needs, and it preserves cell morphology well, which is why it is the usual choice when the downstream goal is DNA. Heparin, which §2.1.1 does not list for whole blood, is a well-known inhibitor of PCR. A bank that accepts heparinized tubes for genomic work therefore has to record that fact and expect extra clean-up steps. The anticoagulant belongs in the sample record even though the enumeration value is the same.

The hemolysis threshold needs careful reading. A hemolysis index is produced by clinical chemistry analyzers that measure free hemoglobin spectrophotometrically, and the scale depends on the analyzer and its manufacturer. The number 50 is therefore the spec's threshold, not a universal unit-bearing constant. A bank should state in its SOP which analyzer produced the index and how the spec's value maps onto that instrument's scale. The spec also gives "up to 10 years" as the storage duration for whole blood. Treat this as the standard's planning horizon rather than a demonstrated shelf life for any particular analyte.

Plasma — PLASMA

Plasma is the cell-free liquid fraction of anticoagulated blood. §2.1.2 is specific: centrifuge at 1,500 g for 10 minutes within 2 hours of collection, store at −80 °C in cryovials, and aliquot into 0.5–1.0 mL portions. The two-hour clock is the most important number in the whole blood section. While cells remain in contact with plasma they keep metabolising glucose, releasing intracellular content and, in the case of platelets, activating. Every minute of delay changes the proteome and metabolome being banked. For that reason the collection time and the processing time must both be recorded; Chapter 5 shows where they live in the lifecycle protocol.

The aliquot size is a design decision about freeze-thaw exposure. A 0.5 mL portion is usually enough for one or two assays. A researcher who needs a small volume can then take one vial without thawing the whole collection. The spec's quality criterion for plasma is total protein 60–80 g/L. This range matches the familiar physiological range for total protein in blood. A value far outside it points to dilution, a labeling error or a processing fault rather than to a genuine donor phenotype, and it should prompt investigation.

Serum — SERUM

Serum differs from plasma in one respect that has large consequences: the blood is allowed to clot first. §2.1.3 specifies 30–60 minutes of clotting at room temperature, then 1,500 g for 10 minutes, then storage at −80 °C. Clotting consumes fibrinogen and releases platelet contents, so serum and plasma from the same donor are not interchangeable for many assays. That is the reason the enumeration keeps them apart. The clotting time actually used can be stored in BloodQuality.clottingTime. Under-clotting leaves fibrin strands that block pipette tips, and over-long clotting lets cells keep metabolising.

The spec adds "stability > 20 years for most analytes". This is the standard's stated guidance. It is not a figure the book can support with a single source, and stability is analyte-specific. A study that depends on a labile analyte should check the literature for that analyte rather than relying on a blanket horizon.

Buffy coat — BUFFY_COAT

After centrifugation, whole blood separates into red cells at the bottom, plasma at the top and a thin whitish interface of leukocytes and platelets: the buffy coat. §2.1.4 describes it as a source for DNA extraction suitable for genomic studies, stored at −80 °C or in liquid nitrogen, with a spec cell count of 1–5 × 10⁶ cells/mL. Harvesting the buffy coat from the same tube that yields plasma is an efficient way to bank both a circulating-protein sample and a germline DNA source from one draw. The buffy coat is not a purified cell population, however. A study that needs living, functional lymphocytes should use PBMCs (Section 2.4).

Blood quality indices

The SDK's BloodQuality interface groups the measurements that apply to all blood derivatives: hemolysisIndex, icterusIndex, lipemiaIndex, totalProtein (g/L) and clottingTime. The first three are the so-called HIL indices that automated analyzers report. They reflect red-cell rupture, bilirubin and lipid turbidity, and each one interferes with particular optical assays. All fields are optional, because not every bank runs every measurement on every sample. The SDK does not define units for the three indices, which is consistent with their analyzer dependence. A bank should document the method so that a value of, say, 30 in one site's data means the same thing as 30 in another's.

2.2 Nucleic Acids

Extracted DNA and RNA are derivatives, not primary specimens. In WIA-BIO-019 terms they are usually aliquots whose parentSampleId points at a whole blood, buffy coat, tissue or saliva record. Chapter 3 covers that genealogy. Here the focus is on what makes an extract good.

DNA — DNA

§2.2.1 lists column-based or organic extraction, an optimal concentration of 50–200 ng/µL and storage from −20 °C to −80 °C. §6.3.1 adds the quality metrics:

• A260/A280 of 1.8–2.0. Nucleic acids absorb maximally near 260 nm and proteins near 280 nm. A ratio near 1.8 is the conventional mark of protein-free DNA, and a markedly lower value suggests protein or phenol carry-over.
• A260/A230 of 2.0–2.2. Many extraction reagents, including guanidine salts and phenol, absorb near 230 nm. A low ratio signals chemical carry-over that can inhibit enzymes downstream even when the A260/A280 looks fine.
• Integrity. The spec asks for no smearing on gel electrophoresis and fragments longer than 10 kb for high-quality genomic DNA. The advanced options are qPCR amplification of long fragments and an automated DNA Integrity Number (DIN).
• Concentration. This is measured by spectrophotometry or fluorometry. Fluorometric dyes bind double-stranded DNA specifically, so they are less inflated by RNA or free nucleotides than absorbance.

The SDK's DNAQuality interface carries concentration (ng/µL), a260_280, a260_230, an optional din and an optional fragmentSize in base pairs. The spec also says DNA is stable for "50+ years at −20 °C", and Appendix B repeats the claim as a maximum duration. As with serum, this is the standard's stated guidance. DNA is indeed the most robust of the banked analytes, but the book does not treat a specific number of decades as established fact.

RNA — RNA

RNA is the most fragile analyte in routine banking. The reason is not the molecule's chemistry alone. RNases are abundant on skin, in dust and in most tissues, and they are notoriously difficult to inactivate. §2.2.2 therefore makes −80 °C mandatory and calls RNase-free technique essential. The spec's extraction options are TRIzol-type organic extraction or columns. Its stated stability is "5 years at −80 °C".

The principal RNA metric is the RNA Integrity Number (RIN), a 1–10 score computed from an electrophoretic trace of ribosomal RNA peaks. §6.3.2 gives interpretation bands and application thresholds:

Secondary metrics are a 28S/18S ribosomal ratio of about 2.0 on a gel and an A260/A280 of 1.9–2.1. Pure RNA sits slightly above DNA on the absorbance ratio. The SDK's RNAQuality mirrors these with concentration, rin, a260_280 and optional ratio_28s_18s. Note what the table does not say. A RIN threshold is a fitness-for-purpose rule, not a pass/fail property of the sample. The same extract can be accepted for a qRT-PCR request and declined for an RNA-seq request. That is why the quality record should hold the measured RIN rather than a pre-computed "OK" flag.

2.3 Tissue: Fresh-Frozen versus FFPE

Surgical and biopsy tissue can be preserved in two fundamentally different ways. The choice decides which research questions the specimen can answer later. WIA-BIO-019 models the two routes as separate types, TISSUE_FRESH_FROZEN and TISSUE_FFPE.

Fresh-frozen tissue

§2.3.1 calls for snap freezing in liquid nitrogen within 30 minutes of excision, pieces of 5–10 mm³, and storage at −80 °C or in the vapor phase of liquid nitrogen. Snap freezing works by crossing the temperature range in which large ice crystals grow so quickly that water is locked in small crystals or a glass-like state. Small pieces freeze faster because heat leaves through their surface. That is why the standard fixes a size window. The 30-minute limit addresses warm ischemia and cold ischemia: once the blood supply is cut, tissue begins to degrade RNA and alter phosphorylation states, and these changes do not wait for the pathology report.

Fresh-frozen tissue is the preferred source for high-quality RNA, native proteins and metabolomics. Its weaknesses are cost, which rises with every year in a −80 °C or LN₂ slot, and morphology, because ice-crystal artefacts make frozen sections harder to read than paraffin sections.

FFPE tissue

Formalin fixation followed by paraffin embedding is the backbone of diagnostic pathology. §2.3.2 specifies fixation in 10% neutral buffered formalin (NBF) for 6–24 hours, then dehydration, clearing and wax infiltration, and storage at room temperature in the dark and dry. The 1:10 tissue-to-fixative volume ratio in common practice exists because formalin penetrates slowly: too little fixative or too short a time leaves the centre of the block under-fixed. Over-fixation, on the other hand, creates more cross-links, which later hampers antigen retrieval for immunohistochemistry (IHC) and fragments nucleic acids.

The spec's duration for FFPE is "decades", with the explicit caveat that DNA and RNA quality degrades. The trade-off is summarised below.

Many banks keep both: a fresh-frozen piece for molecular work and the paired FFPE block from the diagnostic workflow. In WIA-BIO-019 these are two records. They share a donor and can be tied to one collection event, but each has its own type, condition and quality fields.

2.4 Living Cells

Cells that must still be alive after thawing are the most demanding category in a biobank. The goal is no longer to keep molecules intact. It is to keep a membrane-bounded, metabolically active system from destroying itself during freezing and thawing.

PBMCs — PBMC

Peripheral blood mononuclear cells (lymphocytes and monocytes) are isolated by density-gradient centrifugation, which the spec names as Ficoll. §2.4.1 specifies freezing in 10% DMSO in fetal bovine serum or albumin, cooling at 1 °C/min to −80 °C, and then storage in liquid nitrogen at −196 °C. DMSO is a permeating cryoprotectant. It enters the cells and replaces part of the water, which reduces both intracellular ice formation and the osmotic shock as extracellular ice concentrates the surrounding solutes. The controlled 1 °C/min rate is a compromise. Faster cooling traps water inside cells, where it forms lethal ice, and slower cooling dehydrates them excessively. The transfer to liquid nitrogen is needed because −80 °C is warm enough for slow damage to continue over long periods.

The standard gives two post-thaw viability figures. §2.4.1 says "> 90% post-thaw when properly frozen", and §6.3.4 sets the general acceptance thresholds for cells. The consistent reading is that 90% is the target of a well-run process, while the §6.3.4 figures are the minimum to accept a vial.

Cell lines — CELL_LINE

§2.4.2 asks for regular subculture, freezing at an early passage, 10% DMSO or glycerol as cryoprotectant, storage in liquid nitrogen, and three kinds of testing: mycoplasma, sterility and authentication. Early passage matters because cell lines drift genetically and phenotypically with every passage. Mycoplasma is singled out because it is invisible to routine light microscopy, does not cloud the medium and yet changes cell behaviour. Authentication is necessary because cross-contamination and misidentification of cell lines are a documented, recurring problem in research. For human lines it is commonly done by short tandem repeat (STR) profiling. The SDK's CellQuality.contamination object records three booleans (mycoplasma, bacteria, fungi). It has no field for authentication results, so a bank must keep those elsewhere and link them.

Post-thaw quality metrics

§6.3.4 gives the following viability and recovery criteria, and CellQuality carries the corresponding fields (cellCount, viability, viabilityMethod, postThawRecovery).

The thresholds differ by method for a reason. Trypan blue relies on an operator's judgement at the microscope, while flow cytometric dyes distinguish membrane-compromised cells more sharply, so the same population can score differently. That is why viabilityMethod is a required field: a viability number without its method cannot be compared.

Recovery and viability also answer different questions. A vial can show 95% viability among the cells that survived and still have lost half its cells during thaw and washing. Recovery captures that loss. The spec also defines how the cryoprotectant is removed after thaw: dilute gradually to about ten times the volume, centrifuge at 300 g for 5 minutes, and resuspend in fresh medium. Adding the diluent gradually limits osmotic swelling as DMSO leaves the cells. Working promptly limits DMSO's toxicity to cells at warmer temperatures.

2.5 Other Biospecimens

Urine, saliva and stool are increasingly important because they can be collected without a needle, often by the donor at home. That convenience moves pre-analytical control out of the clinic. The metadata recorded at collection therefore becomes the main quality evidence.

Urine — URINE

§2.5.1 specifies clean-catch midstream collection, optional protease inhibitors, storage at −80 °C for metabolomics and aliquots of 10–50 mL. Midstream collection reduces contamination from the urethra and skin. Protease inhibitors protect urinary proteins and peptides, but they add compounds that may interfere with some analyses. The decision should therefore be made per collection and recorded, not applied silently. The spec's stated stability of "5 years for most metabolites" is guidance rather than a guarantee for any particular metabolite.

Saliva — SALIVA

§2.5.2 allows passive drool or swab collection, a volume of 2–5 mL, storage from −20 °C to −80 °C, and lists DNA extraction and microbiome studies as applications. The spec gives a DNA yield of 1–10 µg per sample. Saliva DNA comes from buccal epithelial cells and leukocytes, but also from oral bacteria, so a high total DNA yield does not guarantee a high human DNA yield. The collection method affects both volume and bacterial load, and should be captured alongside the sample.

Stool — STOOL

§2.5.3 specifies a sterile container, a nucleic-acid stabilizing buffer (the spec cites a commercial product "or equivalent"), storage at −80 °C, aliquots of 200–500 mg, and microbiome and metabolomic applications. Without stabilization, the microbial community keeps growing and dying after defecation, and the profile shifts away from what was in the gut. The stabilizer used, and the time from collection to stabilization or freezing, are key metadata. Two stool samples with the same SampleType but different stabilizers may not be comparable in a microbiome analysis.

2.6 Storage Systems in Practice

§4 describes the five storage conditions as equipment classes with their own specifications, safety needs and maintenance. The enumeration values in the SDK carry a comment with the nominal temperature of each: −196 °C, −80 °C, −20 °C, 4 °C and 20–25 °C.

Liquid nitrogen — LIQUID_NITROGEN

§4.1.1 distinguishes the liquid phase at −196 °C from the vapor phase at −150 °C to −190 °C, and gives tank capacities from 100 to 50,000 samples. The advantages the spec lists are the ultra-low temperature, indefinite storage for cells and no dependence on electricity if the tank is kept filled. The disadvantages are cost, the need for regular refilling, safety hazards and a cross-contamination risk in the liquid phase. That last point is the reason many banks prefer vapor-phase storage. Liquid nitrogen can carry micro-organisms, and it can enter a poorly sealed vial and then expand violently when the vial warms.

The safety requirements in §4.1.2 follow from basic physics. Liquid nitrogen expands to roughly 700 times its volume as it evaporates, and it displaces oxygen in an enclosed room without any warning smell. Oxygen monitors, ventilation, cryogenic gloves and face shields, low-level alarms, emergency fill procedures and staff training on cryogenic hazards are therefore requirements, not recommendations. Note a modelling gap here. StorageCondition has a single liquid_nitrogen value with the comment −196 °C, and cannot say whether a vial sits in the liquid or the vapor phase. A bank that uses both should record the phase in the location data or the storage-unit description.

Ultra-low temperature freezers (−80 °C) — ULTRA_LOW_FREEZER

The ULT freezer is the default home for plasma, serum, DNA, most RNA and much tissue. §4.2.1 specifies −80 °C ± 5 °C, a capacity of 300–500 two-inch boxes, LN₂ or CO₂ backup systems and 24/7 temperature logging. The spec lists no need for refilling, easier access and lower running cost than LN₂ as advantages. It lists power dependency, faster degradation than LN₂ and a 10–20-year equipment lifespan as disadvantages. (The spec also quotes a power figure. The book omits it because it is implausibly high for a single −80 °C unit.) §4.2.2 sets the maintenance calendar:

§4.2.2 also lists the required alarms: high and low temperature, power failure, door ajar, and remote monitoring over the network. Chapter 5 turns these alarms into an excursion protocol with thresholds and response steps.

Standard freezers (−20 °C) — STANDARD_FREEZER

§4.3 limits −20 °C storage to long-term DNA, some enzymes and proteins, and short-term holding. It explicitly rules out RNA, notes that proteins degrade over time at this temperature, and warns that frost-free models cause freeze-thaw cycles. Manual-defrost models are the stated best practice. The warning about frost-free units is one of the most useful single lines in the standard, and Section 2.7 returns to it.

Refrigerated storage (4 °C) — REFRIGERATED

§4.4 limits refrigerated storage to short-term holding "< 72 hours", some fixed samples and transport containers, with continuous temperature logging. The same section also says "days to weeks maximum". The two statements are compatible only if the 72-hour figure applies to unfixed biological material awaiting processing and the longer figure to fixed or stabilized material. A bank's SOP should say which applies to which sample type rather than leave it to interpretation.

Ambient storage — AMBIENT

§4.5 covers FFPE blocks, dried blood spots and some fixed specimens at 15–25 °C, below 60% relative humidity, in the dark and free of dust. Humidity matters because paraffin blocks and filter-paper cards absorb moisture, and moisture enables mould and chemical degradation. Two small inconsistencies are worth knowing. The SDK comment for AMBIENT says 20–25 °C and the README table does the same, while §4.5 says 15–25 °C. Dried blood spots are named as an ambient application, but there is no dedicated SampleType for them, so they fall under other. An implementation should take the spec's §4.5 range as the monitoring window, because it is the normative text.

2.7 Pitfalls That Quietly Ruin Samples

// Book example: normalising incoming values before building CellQuality
const cellQuality = {
  cellCount: 5.0e6,              // cells/mL
  viability: 92,                 // percent, 0–100
  viabilityMethod: 'flow_cytometry',
  postThawRecovery: 0.78,        // fraction, 0–1
  contamination: { mycoplasma: false, bacteria: false, fungi: false },
};
if (cellQuality.viability <= 1) {
  throw new Error('viability looks like a fraction; expected percent');
}
if (cellQuality.postThawRecovery > 1) {
  throw new Error('postThawRecovery looks like a percent; expected fraction');
}

Summary

WIA-BIO-019 classifies each specimen by what it is (SampleType, 14 values), where it is kept (StorageCondition, 5 values) and how its quality is measured (five quality interfaces). The numbers attached to each type are mostly well-grounded laboratory conventions: a two-hour window for plasma, snap freezing within 30 minutes, A260/A280 of 1.8–2.0, RIN thresholds by application, 10% DMSO and 1 °C/min for cells, and a maximum of three freeze-thaw cycles for proteins. The weaker parts are the blanket stability durations, the viability criterion for fresh-frozen tissue, and small inconsistencies between the spec, the SDK comments and the README. An implementer should record raw measurements together with their methods and enforce type-to-condition rules in policy. Chapter 3 moves from this scientific taxonomy to the data model that carries it: the Sample record, its location hierarchy and the two integrity metrics.

Review Questions

1. Why does WIA-BIO-019 keep PLASMA and SERUM as separate types even though both are stored at −80 °C? Name two processing differences.
2. A DNA extract shows A260/A280 = 1.85 and A260/A230 = 1.3. What does this suggest, and would you release it for a PCR-based study?
3. An RNA aliquot has RIN 6.5. Which requests in §6.3.2 can it satisfy, and why should the record store 6.5 rather than a pass/fail flag?
4. Explain why the "cell viability > 80%" criterion is a poor acceptance test for snap-frozen tissue. What would you measure instead?
5. List the ULT maintenance tasks in §4.2.2 with their frequencies, and give one failure each task is meant to prevent.
6. Which information about liquid-nitrogen storage cannot be expressed by the StorageCondition enumeration alone, and where would you record it?