A mid-level course for future biochemists and biological technicians, about growing cells and turning them into products. The first module covers cell culture: media and growth conditions, passaging and counting cells with a hemocytometer, and recognising and responding to contamination, including mycoplasma and misidentified cell lines. The second module covers bioprocessing: microbial growth curves and growth rates, how bioreactors control temperature, pH, dissolved oxygen and feeding during scale-up, and the downstream steps that purify a product such as a monoclonal antibody. The third module covers GMP manufacturing: working to a batch record with good documentation practice, equipment qualification (IQ, OQ, PQ) and cleanroom gowning, and reviewing batch data in SQL to flag out-of-specification results. Python and SQL activities are fill-in-the-blank and use only the standard library and SQLite. Every activity is a paper, data or kitchen-safe exercise: nothing here involves cells, microbes, culture reagents or hazardous chemicals at home. Real cell culture and GMP work need supervised training at the right biosafety level in a teaching or employer lab, a degree with lab coursework, and the employer's own GMP training.

Culture cells aseptically, count and passage them, control contamination, model growth, understand bioreactors and purification, and work under GMP with validated systems and batch data.

Cell Culture & Bioprocessing

Find a free product information sheet or formulation table for DMEM on a supplier's website. Note the glucose level, whether it contains L-glutamine and sodium pyruvate, and the sodium bicarbonate concentration.

Calculate the volumes to make 500 mL of complete medium with 10 percent FBS, 1 percent penicillin-streptomycin (100x) and 1 percent L-glutamine (100x). Then repeat for 250 mL with 5 percent FBS.

Write the label you would put on the bottle: contents, final concentrations, lot numbers (make them up), date made, expiry, initials.

Design a culture log table with columns for date, cell line, passage number, medium lot, serum lot, confluence or count, split ratio, observations, incubator ID and initials.

Fill in three days of made-up entries for one flask, including a day when the medium looked slightly yellow and what you did about it.

Design an incubator log with temperature, CO2 percent, water pan check and initials, and note how often your lab would verify CO2 with an independent meter.

If your teaching lab has a cell culture room, ask your instructor to walk you through how they prepare and label medium and use the biosafety cabinet.

## Why it matters

Cultured cells are grown outside the body for research, drug testing, vaccine production and biologics such as monoclonal antibodies. They only behave consistently if their environment is consistent: the same medium, the same supplements, the same temperature, gas and humidity, handled the same way. Many "irreproducible" experiments trace back to a changed serum lot or an incubator that drifted. As a technician you prepare media, keep cultures fed and record every condition.

## What is in culture medium

A **basal medium** (such as DMEM, RPMI 1640 or MEM) supplies:

- **Salts** that set osmolality (roughly 260 to 340 mOsm/kg for most mammalian cells).
- **Glucose** (energy) and **amino acids**, including **L-glutamine**, which breaks down in solution, so it is often added fresh or replaced by a stable dipeptide form.
- **Vitamins** and trace elements.
- A **buffer**, usually sodium bicarbonate, which works with the incubator's CO2.
- Often **phenol red**, a pH indicator: red-pink near pH 7.4, yellow when acidic (cells overgrown or bacteria), purple when too alkaline (for example a flask left out of the incubator).

Common **supplements**:

| Supplement | Purpose |
|---|---|
| Fetal bovine serum (FBS), often 10 percent | Growth factors, hormones, attachment factors; varies lot to lot |
| L-glutamine (if not in the basal medium) | Amino acid that degrades over weeks |
| Penicillin-streptomycin (optional) | Antibiotics; many labs avoid routine use because it can hide poor technique |
| Specific growth factors | For defined, serum-free media used in biomanufacturing |

Biomanufacturing increasingly uses **chemically defined, serum-free** media, because serum is variable and is a possible source of contaminants.

## Worked example: making complete medium

You need 500 mL of DMEM with 10 percent FBS, 1 percent penicillin-streptomycin (supplied at 100x) and 1 percent L-glutamine (supplied at 100x, 200 mM).

```text
FBS:          10% of 500 mL = 50 mL
Pen-strep:    500 mL / 100  = 5 mL
L-glutamine:  500 mL / 100  = 5 mL
Basal DMEM:   500 - 50 - 5 - 5 = 440 mL
```

Many labs simply add supplements to a 500 mL bottle and accept a slightly larger volume; your SOP says which. Either way, label the bottle with the contents, date, your initials and lot numbers, and the expiry your lab assigns (often a few weeks at 2 to 8 °C).

## Growth conditions

Most mammalian cell lines grow at:

- **37 °C**.
- **5 percent CO2** (some media use 10 percent). The CO2 dissolves into the medium and, with bicarbonate, holds the pH near 7.2 to 7.4. Medium matched to the wrong CO2 level drifts in pH.
- **High humidity** (a water pan in the incubator) to stop evaporation, which would concentrate the medium.

Insect cells, many microbes and some specialised lines need different temperatures and gases. Adherent cells grow attached to treated plastic; suspension cells (such as CHO cells in bioreactors) grow free in shaken or stirred flasks.

## Aseptic technique and the biosafety cabinet

Cell culture work happens in a **Class II biological safety cabinet**, which protects the worker, the culture and the environment with HEPA-filtered airflow. Core habits: disinfect the work surface and items going in (commonly with 70 percent ethanol), work in the middle of the cabinet without blocking the grilles, never pass hands over open vessels, use each sterile pipette once and keep bottles capped. Human cell lines are usually handled at **biosafety level 2**, because they can carry unknown viruses. The CDC's BMBL describes biosafety levels: https://www.cdc.gov/labs/bmbl/index.html (Checked on 2026-10-23.) You learn this hands-on in a teaching or employer lab with your instructor, after institution training.

## Keeping a culture log

For each culture, record: cell line and source, passage number, date, medium and lot numbers (especially serum), split ratio or seeding density, confluence or cell count, appearance, and the incubator used. Incubator CO2, temperature and water level are logged too, and the incubator's CO2 reading is checked periodically with an independent meter.

## Common mistakes

- Using medium made up for 10 percent CO2 in a 5 percent incubator.
- Changing the serum lot without testing or recording it.
- Letting the incubator water pan run dry.
- Relying on antibiotics instead of aseptic technique.
- Unlabelled bottles of "complete medium" of unknown age.

## Quick recap

- Medium = basal medium (salts, glucose, amino acids, vitamins, bicarbonate buffer) plus supplements.
- Phenol red shows pH: yellow acidic, purple alkaline.
- Typical conditions: 37 °C, 5 percent CO2, humidified.
- Work aseptically in a biosafety cabinet, and log lots, passage numbers and conditions.

## Try it

Open the **Do** tab to plan a batch of complete medium and design a culture log, on paper.


Understand what goes into cell culture medium and the incubator conditions cells need, and keep a culture log.

Media and growth conditions

Introduction to Cell Culture: Cell Culture Basics Training

1) Cell Culture Tutorial - An Introduction

## Why it matters

Cells in culture keep dividing until they run out of space or nutrients. Left too long, adherent cells become over-confluent and change behaviour, and suspension cultures exhaust their medium. **Passaging** (also called splitting or subculturing) moves a fraction of the cells into fresh medium and vessels. To do this reproducibly, and to seed experiments with the same number of cells every time, you count cells. Getting the count wrong by a factor of two or ten quietly ruins experiments and production runs.

## When to passage

- **Adherent cells** are usually passaged at about 70 to 90 percent **confluence** (the fraction of the surface covered), judged under the microscope, before they fill the flask.
- **Suspension cells** are passaged when they reach a set density, for example a maximum cells/mL in the cell line's datasheet.
- Every passage increases the **passage number**. Cell lines can drift genetically over many passages, so labs work within a defined range and go back to frozen stocks from a **cell bank**.

## How adherent cells are passaged (overview)

1. Remove the old medium.
2. Rinse with a calcium- and magnesium-free salt solution, which loosens cell attachments.
3. Add a dissociation enzyme (such as trypsin) briefly at 37 °C until cells round up and detach.
4. Add medium containing serum (or an inhibitor) to stop the enzyme.
5. Mix gently to break clumps, then take a sample to **count**.
6. Seed the right number of cells into new vessels with fresh medium, label them (cell line, passage number, date, initials) and return them to the incubator.

You do this only in a biosafety cabinet in a teaching or employer lab, after training and under supervision.

## Counting with a hemocytometer

A **hemocytometer** is a thick glass slide with an etched grid. Each **large square** is 1 mm x 1 mm, and with the coverslip the chamber is 0.1 mm deep, so each large square holds **0.1 uL**, or 1/10,000 of a mL.

To tell live from dead cells, the sample is mixed with **trypan blue**: live cells with intact membranes exclude the dye and look bright; dead cells take it up and look blue.

```text
cells/mL = (cells counted / number of large squares) x dilution factor x 10,000
viability (%) = live cells / (live + dead cells) x 100
```

**Worked example.** You mix 10 uL of cells with 10 uL of trypan blue (dilution factor 2) and count four large squares: 52, 48, 55 and 45 live cells (mean 50), and 10 dead cells in total.

```text
live cells/mL = 50 x 2 x 10,000 = 1,000,000 cells/mL (1.0 x 10^6)
viability     = 200 / (200 + 10) x 100 = 95.2 %
```

To seed a flask with 5 x 10^5 cells: volume = 500,000 / 1,000,000 = **0.5 mL** of suspension, topped up with fresh medium to the flask's working volume.

Counting rules that keep counts consistent: count cells touching the top and left borders of a square but not the bottom and right; aim for roughly 20 to 100 cells per square (dilute further if there are more); and count at least four squares. Many labs now use automated counters, which apply the same arithmetic.

## Split ratios

For routine maintenance, labs often use a **split ratio** instead of an exact count. A 1:5 split puts one fifth of the cells into a new vessel of the same size. Experiments that compare conditions almost always need an actual count, so every well starts with the same number of cells.

## Common mistakes

- Forgetting the trypan blue dilution factor (a 2x error).
- Counting too few cells, so random error dominates.
- Letting cells settle in the tube before taking the counting sample.
- Leaving trypsin on too long, which damages cells and lowers viability.
- Not recording the passage number, so a culture drifts outside its validated range.

## Quick recap

- Passage adherent cells at about 70 to 90 percent confluence; track passage number.
- Hemocytometer: cells/mL = mean count per large square x dilution x 10,000.
- Trypan blue: live cells exclude it; viability = live / total x 100.
- Seeding volume = cells wanted / cells per mL.

## Try it

Open the **Do** tab. A Python counter is written except for the cells-per-mL formula.


Know when and how cells are passaged, count them with a hemocytometer and compute seeding volumes.

Passaging and counting

How to Passage Cells: Cell Culture Basics Training

Counting Cells with a Hemocytometer

## Why it matters

Contamination is the most common way cell culture goes wrong. Some contaminants are obvious overnight; others, such as mycoplasma or a different cell line, can sit unnoticed for months and invalidate every result produced from those cells. In biomanufacturing a contamination event can mean losing a batch worth a great deal of money and investigating the whole facility. A technician who recognises contamination early, and responds by the procedure rather than by instinct, protects the whole lab.

## The main types

| Contaminant | What you see | Notes |
|---|---|---|
| **Bacteria** | Medium turns cloudy and often yellow within a day or two; tiny moving dots between cells under the microscope | Fast growing; sometimes a film on the surface |
| **Yeast** | Cloudy medium; round or oval particles, sometimes budding, often in chains | pH may change slowly at first |
| **Mould (fungi)** | Thin filaments (hyphae), later fuzzy clumps visible by eye | Spores spread through air and incubators |
| **Mycoplasma** | Usually nothing visible: no cloudiness, no obvious pH change | Very small bacteria without a cell wall; alter cell metabolism and gene expression; not killed by penicillin |
| **Viruses** | Usually nothing visible | Important in biomanufacturing; tested with specific assays |
| **Cross-contamination** | Cells look and behave like "your" line but are not | Another cell line has taken over; HeLa is the classic culprit |

## Detecting the invisible ones

- **Mycoplasma** is tested routinely (for example monthly, and always for new lines entering the lab) using PCR, enzyme-based luminescence kits or DNA staining. Bought or shared lines go into quarantine until they test clean.
- **Cell line identity** is confirmed by **STR profiling** (short tandem repeat analysis) for human lines, compared against reference databases. The International Cell Line Authentication Committee (ICLAC) keeps a register of known misidentified lines: https://iclac.org/databases/cross-contaminations/ (Checked on 2026-10-23.)
- In GMP manufacturing, cell banks and production batches are tested for sterility, mycoplasma and adventitious viruses before release.

## Prevention

- **Aseptic technique** in a certified biosafety cabinet, every time.
- **One cell line at a time** in the cabinet, with cleaning between lines, so lines cannot mix.
- **Separate, labelled media** bottles for each line, never shared.
- Clean incubators and water pans on a schedule, and keep the lab tidy.
- Work from a tested **master and working cell bank**, and keep passage numbers low.
- Minimise routine antibiotics, which can mask low-level contamination.

## Responding to contamination

When you suspect contamination, follow your lab's procedure. Typically:

1. **Do not open the vessel** in the cabinet with other cultures, and do not try to "rescue" it unless the culture is irreplaceable and your supervisor approves a specific procedure.
2. **Tell your supervisor** and record what you saw, with photographs if possible.
3. **Decontaminate and discard** the culture as your institution's biosafety procedure requires (for example by disinfectant treatment or autoclaving).
4. **Check everything that was shared:** other cultures from the same incubator, media and reagent bottles used with it, and the cabinet.
5. **Clean** the cabinet and incubator as required.
6. **Investigate the cause** (technique, a reagent lot, incubator, a new cell line) and record corrective actions, using the deviation and CAPA approach from the lab data course.

**Worked example.** Monday you see cloudy, yellow medium in one of three flasks of the same line, with moving dots under the microscope. The other two flasks look normal. Bacterial contamination of one flask is likely. You report it, discard the flask as required, check the other two flasks and the shared medium bottle, and clean the cabinet. Because all three were fed from the same medium bottle on Friday, you take a sample of that bottle for testing and quarantine it until the result is back.

## Safety and training

Contaminated cultures, and human or animal cell lines in general, are biohazards. Handling, disinfection and waste rules depend on your institution and biosafety level, and you learn them through institution training under supervision. This course never asks you to handle cultures at home.

## Common mistakes

- Assuming no cloudiness means no contamination (mycoplasma and cross-contamination are invisible).
- Trying to treat a contaminated flask with extra antibiotics and carrying on.
- Opening the contaminated flask in the shared cabinet.
- Skipping quarantine for a new line from a collaborator.
- Not recording the event, so a pattern across weeks is missed.

## Quick recap

- Bacteria, yeast and mould are usually visible; mycoplasma, viruses and cross-contamination usually are not.
- Test mycoplasma routinely and authenticate lines (STR); check the ICLAC register.
- Prevent with aseptic technique, one line at a time, separate media and tested cell banks.
- Respond by the procedure: contain, report, discard safely, check shared items, clean, investigate, document.

## Try it

The **Do** tab has a quiz with contamination scenarios.


Recognise bacterial, fungal, mycoplasma and cross-contamination in cell culture and respond correctly.

Contamination control

Cell Culture Contamination and Prevention

The dangers of mycoplasma contamination

Prepare and log media and growth conditions, count and passage cells, and identify and respond to contamination.

Cell Culture

## Why it matters

Bacteria and yeast are the workhorses of many bioprocesses: they make insulin, enzymes, plasmid DNA and many other products. Knowing how fast a culture grows tells you when to induce product expression, when to feed, when to harvest and whether something is wrong (a slow culture can mean a bad medium lot, wrong temperature or contamination). Growth rates also let you compare strains and media objectively instead of saying "it looked a bit slower".

## The phases of a batch growth curve

Inoculate microbes into fresh medium in a closed flask or bioreactor and measure them over time, and you see four phases:

1. **Lag phase:** cells adjust to the new medium; little increase in number.
2. **Exponential (log) phase:** cells divide at a constant rate; the population doubles every fixed interval.
3. **Stationary phase:** a nutrient runs out or waste builds up; growth and death balance.
4. **Death (decline) phase:** viable cells decrease.

Growth is usually followed by **optical density at 600 nm (OD600)**, which measures light scattered by the cells. OD is only proportional to cell concentration over a limited range (often up to about 0.4 to 1.0 depending on the instrument), so dense samples are diluted before reading and the reading multiplied back. Viable counts (colony-forming units per mL) and dry weight are other measures.

## The maths of exponential growth

In the exponential phase:

```text
X(t) = X0 x e^(mu x t)
ln X(t) = ln X0 + mu x t
```

- **X** is biomass or OD, **t** is time, and **mu** is the **specific growth rate** (per hour).
- Taking the natural log turns the curve into a **straight line**, and its slope is mu. That is why growth curves are plotted on a log scale.
- **Doubling time** td = ln(2) / mu.

**Worked example.** OD600 is 0.050 at 2.0 h and 0.302 at 4.0 h, both in the exponential phase.

```text
mu = (ln 0.302 - ln 0.050) / (4.0 - 2.0) = (-1.197 - (-2.996)) / 2.0 = 0.90 per hour
td = 0.693 / 0.90 = 0.77 h, about 46 minutes
```

Using only two points is sensitive to noise. Better practice is to fit a straight line (linear regression) through ln(OD) for all points in the exponential phase, which is what the activity does.

## Choosing the exponential phase

Plot ln(OD) against time. The exponential phase is the straight section. Leave out:

- the **lag** at the start, where the line is flat;
- the **slowing** near stationary phase, where the line bends over;
- any readings outside the instrument's linear range.

Including lag or stationary points underestimates mu. Fitting tools in Python (for example `scipy.optimize.curve_fit` with a logistic or Gompertz model) can fit the whole curve, but a log-linear fit of the exponential window is the standard first step and is easy to check by eye.

## Where it is used in bioprocessing

- **Induction timing:** many recombinant *E. coli* processes add the inducer at a set OD in mid-exponential phase.
- **Fed-batch feeding:** feed rates are set to hold a target growth rate and avoid oxygen limitation or overflow metabolism.
- **Process monitoring:** online sensors (OD probes, off-gas CO2 and O2) track growth in real time in bioreactors.
- **Comparing conditions:** media, temperature and strain changes are judged by mu and final biomass.

Mammalian cells grow much more slowly than bacteria (doubling times of roughly a day), but the same log-linear maths applies to viable cell density.

## Safety and training

Growing microbes is done only in a teaching or employer lab with appropriate containment, sterilisation and waste procedures, under supervision. The activity here uses data only.

## Common mistakes

- Fitting a straight line to raw OD instead of ln(OD).
- Including lag or stationary points in the fit.
- Reading OD above the linear range without diluting.
- Mixing log10 and natural log: if you use log10, mu = slope x ln(10).
- Reporting a doubling time in hours as minutes, or the reverse.

## Quick recap

- Batch growth: lag, exponential, stationary, death.
- In exponential phase ln(OD) is linear in time; the slope is mu.
- Doubling time td = ln 2 / mu.
- Fit only the straight section, within the instrument's linear range.

## Try it

Open the **Do** tab. A Python script fits the growth rate; you add the log transform that makes it work.


Recognise the phases of a growth curve and fit the specific growth rate and doubling time from OD data.

Microbial growth curves

Bacterial Growth Curve

Specific Growth Rate Formula: What Is Specific Growth Rate (μ)?

## Why it matters

A shake flask is fine for a few hundred millilitres. Producing a vaccine or a monoclonal antibody for thousands of patients needs bioreactors of hundreds to thousands of litres, run identically batch after batch. Technicians in upstream manufacturing set up, monitor and sample bioreactors, and must understand what each control loop is doing and why a parameter drifting out of range matters.

## The stirred-tank bioreactor

The most common design is a **stirred tank**: a vessel (stainless steel, or a **single-use** plastic bag inside a support) with:

- an **impeller** that mixes the culture;
- a **sparger** that bubbles air, oxygen, nitrogen or CO2 into the liquid;
- **probes** for temperature, pH and **dissolved oxygen (DO)**, plus often online cell density, and off-gas analysis;
- **ports** for adding base, feeds and antifoam, and for taking samples aseptically;
- a jacket or heating blanket for temperature control.

A control system reads the probes and adjusts the equipment to hold each parameter at its **set point**.

## The main control parameters

| Parameter | Typical mammalian set point | How it is controlled |
|---|---|---|
| Temperature | about 36 to 37 °C (sometimes shifted lower later in the run) | Jacket heating and cooling |
| pH | about 6.9 to 7.2 | CO2 sparging lowers pH; base addition raises it |
| Dissolved oxygen | often 30 to 50 percent of air saturation | Sparging air or O2, agitation speed |
| Agitation | set to mix well without damaging cells | Impeller speed |
| Feeds, glucose | kept in a target range | Scheduled or controlled feed additions |
| Foam | kept low | Antifoam addition |

Microbial fermentations run hotter and much faster, and need far more oxygen, so their agitation and aeration are more intense. Oxygen transfer is often the limit, described by the **kLa** (volumetric mass transfer coefficient).

## Operating modes

- **Batch:** everything is added at the start; cells grow until a nutrient runs out. Simple but low yield.
- **Fed-batch:** concentrated nutrient feeds are added during the run, extending growth and production. This is the standard for most monoclonal antibody production in CHO (Chinese hamster ovary) cells, typically over about two weeks.
- **Perfusion:** fresh medium flows in while spent medium (and often product) is removed, with cells retained by a filter. High cell densities and continuous production, but more complex.

## Scale-up

Production starts from a vial of the **working cell bank**, which is expanded through a **seed train** of progressively larger flasks and bioreactors until there are enough cells to inoculate the production bioreactor.

Scaling up is hard because not everything scales together. In a bigger tank:

- **Mixing time** gets longer, so added base or feed takes longer to disperse, and local pH or nutrient gradients form.
- **Oxygen transfer** and **CO2 removal** change: dissolved CO2 can build up in large tanks.
- **Shear** from the impeller tip and bursting bubbles can damage sensitive cells.
- The **surface-to-volume ratio** falls, so heat removal and aeration behave differently.

Engineers therefore choose which criteria to keep constant, commonly **power input per volume (P/V)**, **kLa**, **impeller tip speed** or **gas flow per volume (vvm)**, and accept changes in others. Scale-down models (small bioreactors that mimic the large one) are used to study problems without running full-size batches.

**Worked example.** A 2 L bench bioreactor runs at 0.1 vvm air (0.1 volume of gas per volume of liquid per minute), which is 0.2 L/min. Keeping vvm constant at 2,000 L means 200 L/min. Keeping superficial gas velocity constant instead would give a very different number, which is why scale-up needs a deliberate choice and verification runs.

## Process data and GMP

In GMP manufacturing, critical process parameters and their ranges are defined during development and validation. Excursions outside the ranges are deviations that must be investigated. Batch data (set points, alarms, samples, operator entries) form part of the batch record you will review in module 3. Real bioreactor operation is learned on the job, under supervision and the employer's training programme.

## Common mistakes

- Treating a DO or pH probe reading as truth without regular calibration and offline checks.
- Ignoring dissolved CO2 build-up at large scale.
- Changing agitation to fix DO without considering shear.
- Breaking sterility at a sampling port.
- Assuming a process that works at 2 L will behave the same at 2,000 L.

## Quick recap

- Stirred-tank bioreactors control temperature, pH, DO, agitation, feeds and foam.
- Batch, fed-batch (most common for mAbs) and perfusion modes.
- Seed train from the working cell bank to the production bioreactor.
- Scale-up changes mixing, oxygen and CO2 transfer, shear and heat; pick scale-up criteria deliberately.

## Try it

The **Do** tab has a quiz on control parameters and scale-up.


Explain what a bioreactor controls, the batch, fed-batch and perfusion modes, and what changes when you scale up.

Bioreactors and scale-up

How a bioreactor works

Bioprocess design and scale-up

## Why it matters

At the end of a production run the product, for example a monoclonal antibody, is a tiny fraction of what is in the bioreactor: cells, cell debris, host-cell proteins, DNA, media components, and possibly viruses. **Downstream processing** turns that mixture into a pure, safe, concentrated drug substance. It is often the most expensive part of manufacturing, and every step costs yield. Technicians run chromatography, filtration and buffer preparation, and need to know what each step is for.

## The usual order

Downstream processing usually follows this logic:

1. **Harvest and clarification:** remove cells and debris.
2. **Capture:** grab the product and concentrate it, removing most impurities in one step.
3. **Virus inactivation:** chemically or physically kill enveloped viruses.
4. **Polishing:** remove remaining trace impurities, aggregates and variants.
5. **Virus filtration:** physically remove viruses by size.
6. **Concentration and buffer exchange:** bring the product to final concentration in its formulation buffer.
7. **Sterile filtration and fill:** filter through 0.2 um filters into containers for storage or further processing.

A typical monoclonal antibody (mAb) platform looks like this:

| Step | Common technology | Removes or achieves |
|---|---|---|
| Clarification | Centrifugation and/or depth filtration | Cells, debris |
| Capture | **Protein A affinity chromatography** | Most host-cell proteins, DNA, media; concentrates the mAb |
| Virus inactivation | Low-pH hold (for example around pH 3.5 to 3.7 for a set time) | Enveloped viruses |
| Polishing | **Cation and/or anion exchange chromatography** (sometimes hydrophobic interaction) | Aggregates, residual host-cell proteins and DNA, leached Protein A, variants |
| Virus filtration | Nanofiltration (pores around 20 nm) | Viruses, including small non-enveloped ones |
| UF/DF | **Ultrafiltration/diafiltration** by tangential flow filtration (TFF) | Concentrates product; exchanges into formulation buffer |
| Final filtration | 0.2 um sterilising-grade filter | Bacteria |

Products made in bacteria differ: *E. coli* often makes proteins inside the cell, sometimes as insoluble inclusion bodies, so downstream starts with cell lysis and may include protein refolding.

## How chromatography separates

- **Affinity:** a ligand binds the product specifically (Protein A binds the Fc region of antibodies). Impurities flow through; the product is eluted, often with low pH.
- **Ion exchange:** proteins bind according to their charge at a given pH, and are eluted by increasing salt or changing pH. Depending on conditions, the product either binds while impurities flow through (bind and elute), or the impurities bind while the product flows through (flow-through mode).
- **Hydrophobic interaction and size exclusion** separate by surface hydrophobicity and size.

## Yield adds up

Each step loses some product, and losses multiply.

**Worked example.** Six steps with yields of 95, 90, 98, 92, 95 and 97 percent:

```text
overall yield = 0.95 x 0.90 x 0.98 x 0.92 x 0.95 x 0.97 = about 0.71 (71 percent)
```

Removing or improving even one step can matter a great deal at manufacturing scale. Process development balances purity against yield and cost.

## What quality checks look for

In-process and release testing measure product concentration, purity (often by size-exclusion HPLC and capillary electrophoresis), **host-cell protein** and **host-cell DNA** levels, aggregates, **endotoxin**, bioburden, and potency. Virus clearance is demonstrated by spiking studies during validation, not tested on every batch. Specifications for these tests are defined in the product's regulatory filing, guided by international guidelines such as ICH Q6B for biotechnological products (ICH quality guidelines: https://www.ich.org/page/quality-guidelines, checked on 2026-10-23).

## Common mistakes

- Thinking of purification as one step rather than orthogonal steps that each remove different impurities.
- Forgetting the cumulative effect of small yield losses.
- Treating a low-pH hold as optional when it is a validated virus-inactivation step with set time and pH.
- Confusing ultrafiltration (concentrates product) with sterile filtration (removes bacteria).
- Holding product between steps longer than validated hold times.

## Quick recap

- Clarify, capture, inactivate viruses, polish, filter viruses, UF/DF, sterile filter.
- mAb platform: Protein A capture, low-pH hold, ion exchange polishing, nanofiltration, TFF, 0.2 um filtration.
- Each step removes different impurities; yields multiply.
- Quality tests cover purity, HCP, DNA, aggregates, endotoxin and potency.

## Try it

The **Do** tab has a quiz on ordering steps and what each removes.


Order the steps that recover and purify a biological product and explain what each removes.

Downstream processing

Bioprocessing Part 2: Separation / Recovery

Bioprocessing Part 3: Purification

Fit microbial growth curves, explain how bioreactors are controlled and scaled up, and order the steps of downstream purification.

Bioprocess

Write a master batch record for "500 g of 10 percent w/w sugar solution": product name, a record number and version, materials (sugar 50.0 g, range 49.5 to 50.5 g; water 450 g, range 445 to 455 g), equipment (scale with an ID you assign), numbered steps, and fields for actual value, performed by and verified by.

Add a step to record the start time and end time, and a final step for appearance (clear, no undissolved sugar).

Issue yourself a copy: write "Batch 001" and the date on it.

Execute it: tare the scale, weigh the sugar and record the actual weight immediately in ink, then the water, then mix until dissolved. Record every value at the time.

Ask a friend or family member to watch the weighing steps and sign as verifier (or role-play both and note it).

Make one deliberate recording error and correct it with GDP: single line, correct value, initials, date, reason.

Write a planned deviation: for example, you mixed for 4 minutes instead of the written 2 minutes. Record what happened, the impact on the product and who you told.

Review your own completed record as QA would: any blank fields, missing signatures or unexplained changes? List them.

## Why it matters

Medicines, vaccines and many biological products must be made under **Good Manufacturing Practice (GMP)**: rules that make sure every batch is made the same validated way, by trained people, with a complete record. A patient cannot test their own medicine, so the record is the evidence. For many biological technicians, GMP is the daily working environment, and the basic skill is executing a batch record correctly. As the saying goes: if it is not documented, it did not happen.

## What GMP covers

GMP regulations and guidelines cover, among other things:

- **Personnel:** qualified, trained people with training records for every procedure they perform.
- **Facilities and equipment:** designed, qualified, cleaned and maintained (next lesson).
- **Materials:** approved suppliers, identity testing, quarantine until release, traceable lot numbers.
- **Procedures:** written, approved SOPs and batch records, followed exactly.
- **Records:** complete, accurate, contemporaneous, reviewed.
- **Quality unit:** an independent quality assurance (QA) function that approves procedures and releases or rejects batches.
- **Deviations, CAPA and change control:** anything unplanned is recorded and investigated; changes are assessed before they are made.

In the US, the core drug GMP regulations are 21 CFR Parts 210 and 211, with additional requirements for biologics in Parts 600 to 680. FDA's overview: https://www.fda.gov/drugs/pharmaceutical-quality-resources/current-good-manufacturing-practice-cgmp-regulations (Checked on 2026-10-23.) In the EU, EudraLex Volume 4 sets GMP: https://health.ec.europa.eu/medicinal-products/eudralex/eudralex-volume-4_en (Checked on 2026-10-23.) Other regions follow similar rules, often harmonised through PIC/S and ICH.

## The batch record

A **master batch record** is the approved template for making one batch of a product: materials and quantities, equipment, step-by-step instructions, process parameters and limits, sampling, and signature fields. For each batch, a controlled copy is issued and becomes the **executed batch record** as operators fill it in.

Executing a step typically means:

1. **Read** the step before doing it.
2. **Do** exactly what it says, with the equipment and materials it names.
3. **Record** the actual values (weights, times, temperatures, lot numbers, equipment IDs) at the time, in permanent ink or the electronic system.
4. **Sign and date** ("performed by").
5. A second trained person **verifies** critical steps (such as weighing or adding an ingredient) and signs "verified by".

## Good documentation practice (GDP)

GDP applies the ALCOA+ principles you learned in the lab data course:

- Write entries **at the time**, never from memory later, and never pre-fill.
- Record **actual** values with units, not "OK" or a tick where a number is required.
- Correct errors with a **single line**, the correct value, initials, date and a reason. No correction fluid, no overwriting.
- Never leave blanks: write "N/A" with initials if a field does not apply, as your procedure allows.
- Use your **own** signature or login; never sign for someone else.
- If something goes differently from the instruction, **stop** and record a **deviation** following your SOP, and tell your supervisor. Do not improvise and carry on.

**Worked example.** Step 12 says "Add 25.0 kg (24.8 to 25.2 kg) of buffer component X, lot to be recorded. Verify." You weigh 25.1 kg, record "25.1 kg", the lot number, the balance ID, the time, sign "performed by", and a colleague who watched the weighing signs "verified by". If you had weighed 25.4 kg, you would not just record it and continue: it is outside the range, so you stop and follow the deviation procedure.

## After the batch

QA reviews the executed batch record, along with test results, deviations and environmental data, before the batch can be **released**. Any gap (a missing signature, an unexplained correction) delays release and can trigger an investigation. You will review batch data yourself in this module's last lesson.

## Training reality

GMP training is given by the employer, specific to its procedures, and repeated regularly. Many community colleges and universities offer biomanufacturing certificates with GMP and aseptic modules. You do not work unsupervised in a GMP area until you are trained and qualified on that procedure.

## Common mistakes

- Filling in several steps at once at the end of the shift.
- Writing a value outside the range and carrying on.
- Ticking "verified" without watching the step.
- Using a stray piece of paper as a "rough record".
- Unsigned or undated corrections.

## Quick recap

- GMP: trained people, qualified equipment, controlled materials, written procedures, complete records, independent QA.
- Master batch record (template) vs executed batch record (this batch).
- Read, do, record actual values, sign, verify critical steps.
- GDP: contemporaneous, single-line corrections with reason, no blanks, no signing for others, stop and record deviations.

## Try it

Open the **Do** tab to write and execute a mock batch record for a kitchen-safe sugar solution.


Explain what GMP requires of people and records, and follow a batch record with good documentation practice.

Working under GMP

What is Good Manufacturing Practice GMP in Pharmaceuticals?

What is Batch Record Review? Explained in 2 minutes! | Free GMP Training video 2025

## Why it matters

GMP does not trust equipment or processes because they look right. It requires documented evidence that they work as intended, every time: this is **validation**. And for products that cannot be terminally sterilised, such as many biologics, the room itself is part of product protection: **cleanrooms** control particles and microbes, and people, the biggest source of contamination, enter them only after gowning in a fixed sequence. Technicians execute qualification protocols, perform environmental monitoring and gown many times a day.

## Equipment qualification: IQ, OQ, PQ

Before a new bioreactor, incubator, autoclave or HPLC is used for GMP work, it is qualified against its **user requirements specification (URS)**:

| Stage | Question | Examples |
|---|---|---|
| **DQ** design qualification | Is the design suitable for the URS? | Review of specifications and drawings |
| **IQ** installation qualification | Is it installed correctly, as specified? | Correct model and parts, utilities connected, calibration certificates, manuals and drawings filed, software version recorded |
| **OQ** operational qualification | Does it operate across its intended ranges? | Temperature control at low, mid and high set points; alarms trigger; door interlocks work |
| **PQ** performance qualification | Does it perform consistently under real conditions? | Temperature mapping of a loaded incubator over days; a sterilisation cycle with a real load and biological indicators |

Each stage has an approved **protocol** written before testing, with acceptance criteria. Results go in a **report**; any failure is a deviation that must be resolved. After qualification, equipment stays in a validated state through calibration, preventive maintenance, periodic review and **change control**: a change (new software version, moved equipment) is assessed and may need requalification.

## Process validation

Qualified equipment is necessary but not enough: the **process** must be shown to make product of the required quality consistently. FDA's process validation guidance describes three stages: **process design** (development defines critical parameters), **process qualification** (including process performance qualification batches at commercial scale) and **continued process verification** (ongoing monitoring of every batch). Cleaning procedures, analytical methods and computerised systems are validated too.

## Cleanrooms

Cleanrooms are classified by airborne particle counts. ISO 14644-1 defines classes such as ISO 5, 7 and 8. EU GMP Annex 1 (in EudraLex Volume 4: https://health.ec.europa.eu/medicinal-products/eudralex/eudralex-volume-4_en, checked on 2026-10-23) uses **grades A to D** for sterile manufacturing:

- **Grade A:** the critical zone where sterile product or open containers are exposed (inside a unidirectional airflow hood or isolator).
- **Grade B:** the background room around Grade A for aseptic work.
- **Grades C and D:** less critical steps, such as preparing solutions.

Key design features: HEPA-filtered air supply, many air changes per hour, a **pressure cascade** (cleaner rooms at higher pressure, so air flows outward toward dirtier areas), **airlocks** between grades, smooth cleanable surfaces, and controlled flows of people and materials.

**Environmental monitoring** checks that the room stays in control: particle counts, settle plates and air samples for microbes, contact plates on surfaces and on operators' gloves and gowns. Results are trended, and excursions are investigated. **Aseptic process simulations (media fills)** run the aseptic process with growth medium instead of product to show that operators and process do not introduce contamination.

## Gowning

People shed skin cells and microbes constantly, so gowning procedures are strict and sequential, moving from "dirty" to "clean":

1. Remove outdoor clothing and jewellery; wash hands.
2. Put on shoe covers or dedicated shoes, crossing a step-over bench.
3. Put on a hood or hair cover and face mask (and goggles as required).
4. Put on the sterile coverall without letting it touch the floor, then sterile boots over it.
5. Put on sterile gloves, often two pairs, sanitised regularly.
6. Check yourself in a mirror for gaps.

The order matters: you cover the head before the coverall, so hair and skin particles fall onto clothing you will cover, not onto the clean suit. Operators are **gowning qualified** (with contact plates taken from the gown) before entering Grade A/B areas, and requalified periodically. Inside, they move slowly and deliberately and never touch critical surfaces.

## Common mistakes

- Running OQ before IQ is approved, or testing without an approved protocol.
- Moving or modifying qualified equipment without change control.
- Letting the coverall touch the floor during gowning.
- Fast movements and talking over open product in Grade A.
- Treating environmental monitoring excursions as noise instead of trends to investigate.

## Quick recap

- IQ: installed correctly; OQ: operates across ranges; PQ: performs consistently in real use. Protocol first, report after.
- Validation covers processes, cleaning, methods and computerised systems; change control keeps the validated state.
- Cleanrooms: HEPA air, pressure cascade, airlocks; EU grades A to D; environmental monitoring and media fills.
- Gowning follows a fixed dirty-to-clean order; operators are gowning qualified.

## Try it

The **Do** tab has a quiz on qualification stages and gowning.


Explain equipment qualification (IQ, OQ, PQ), process validation, cleanroom grades and why gowning is done in a fixed order.

Validation and cleanrooms

Qualification Phases: IQ, OQ, and PQ Explained Simply

Aseptic Gowning for the Cleanroom

## Why it matters

Before a batch of drug substance can be released, someone has to confirm that every required test was done and every result meets its **specification**. A single missed out-of-specification (OOS) result can put an unsafe batch on the market; a sloppy investigation that "tests into compliance" can lead to regulatory action. Batch data increasingly lives in databases (LIMS and manufacturing execution systems), so being able to query it, not just read printouts, makes review faster and more reliable.

## Specifications, OOS and OOT

- A **specification** is the list of tests, methods and acceptance criteria a batch must meet, for example pH 5.8 to 6.4 or endotoxin not more than 0.5 EU/mg. Some limits are two-sided; others have only a lower limit (purity at least 95.0 percent) or only an upper limit (endotoxin).
- An **out-of-specification (OOS)** result falls outside an acceptance criterion.
- An **out-of-trend (OOT)** result is inside specification but unusual compared with previous batches, for example a sudden step change. It is a warning worth investigating, just like a Westgard warning in clinical QC.

Boundary values matter: if the limit is "not more than 0.5" then 0.5 passes and 0.51 fails. Specifications also state how results are rounded before comparison, so follow the written rule.

## What a batch data review checks

1. **Completeness:** every required test has a result, from a valid run, with the right method version.
2. **Compliance:** each result meets its specification.
3. **Trends:** results compared with previous batches (control charts).
4. **Related records:** the executed batch record is complete, deviations are closed or assessed, environmental monitoring is in control, and equipment was qualified and calibrated.
5. **Data integrity:** audit trails show no unexplained changes, deletions or repeated injections.

Only after QA is satisfied is the batch released, and a certificate of analysis (CoA) issued.

## Querying batch data in SQL

Suppose a LIMS has `results(batch_id, test, value, analyst)` and `specs(test, lower_limit, upper_limit, units)`, where a missing limit is stored as `NULL`. Joining them puts every result next to its limits:

```sql
SELECT r.batch_id, r.test, r.value, s.lower_limit, s.upper_limit
FROM results AS r
JOIN specs AS s ON s.test = r.test;
```

To keep only OOS results, add a condition that the value is below the lower limit **or** above the upper limit. In SQL, any comparison with `NULL` is "unknown" rather than true, so a test with no upper limit simply never fails the upper check, which is exactly what we want here. (If you write it the other way round, as "keep rows that are within limits", `NULL` limits would wrongly exclude rows, so be deliberate.)

**Worked example.** pH limits 5.8 to 6.4. Batch B-202 has pH 6.6: 6.6 > 6.4, so it is OOS. Batch B-204 has pH exactly 6.4: not greater than 6.4, so it passes.

## What happens after an OOS

In the US, FDA's guidance *Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production* sets expectations: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/investigating-out-specification-oos-test-results-pharmaceutical-production-level-2-revision (Checked on 2026-10-23.) In outline:

1. **Phase I, laboratory investigation:** the analyst and supervisor promptly check for an obvious, documented laboratory error (wrong dilution, instrument failure, calculation error), preserving the original samples and data. Only a clearly identified lab error can invalidate the result.
2. **Phase II, full-scale investigation:** if no lab error is found, the investigation extends to manufacturing: batch record, deviations, raw materials, equipment and other batches that might be affected.
3. **Retesting and resampling** follow predefined rules. Testing again and again until a passing result appears ("testing into compliance") is not acceptable, and averaging cannot be used to hide an individual OOS value.
4. The conclusion, root cause and CAPA are documented, and QA decides the batch's fate.

## Common mistakes

- Reviewing only the tests that look interesting, not checking completeness.
- Treating a value exactly at the limit as a failure (or a value just past it as a pass) without reading the specification wording.
- Mishandling `NULL` limits in queries, which silently drops or wrongly flags rows.
- Discarding an OOS result as "probably a lab error" without evidence.
- Ignoring OOT results because they are technically in specification.

## Quick recap

- Specifications define tests and acceptance criteria; OOS is outside them, OOT is unusual but inside.
- Review covers completeness, compliance, trends, related records and data integrity before QA release.
- In SQL, join results to specs and filter value < lower OR value > upper; know how `NULL` behaves.
- OOS investigations: Phase I lab check, Phase II full-scale, controlled retesting, documented conclusion; never test into compliance.

## Try it

Open the **Do** tab and write the `WHERE` clause that flags out-of-specification results.


Review batch test results against specifications, flag out-of-specification results in SQL and know what happens next.

Batch data review

Investigation of Out of Specification Results | OOS Investigation

Batch Record Review Process in Pharmaceutical Manufacturing. GMP Batch Release and Disposition.

Follow a batch record correctly, explain IQ/OQ/PQ and cleanroom gowning, and flag out-of-specification batch results in SQL.

Cell Culture & Bioprocessing

About this course

Course content

Cell Culture

Bioprocess

GMP Manufacturing

Where this course leads

Careers that use it