Chromatography methods in life-science labs

Chromatography methods in life-science labs separate molecules by giving each one a different average stay on a stationary phase while a mobile phase carries them forward. The decision this page supports is which mode matches the chemical difference you can actually exploit, and what a peak on a trace is allowed to mean. Preparative protein work with tags and charges continues in affinity and ion-exchange protein purification.

Hardware classes are listed in the scientific instruments catalogue. A method question about a protein campaign can be framed with the custom protein expression and purification reference. That page is an enquiry prompt. It does not say EVRINTH performs the purification. Specifications and solvent constraints go through the quote request.

The separation you are asking for

Start from the difference between the molecule you want and the molecules you have. Charge, size, hydrophobicity and a unique binding partner are different differences. If you cannot name one, a column will still produce a line on a screen, and the line will not answer you. Analytical chromatography asks "what is the pattern, and how much, relative to a standard?". Preparative chromatography asks "which fraction do I keep?". Mixing those questions is how people dry down an analytical injection and call it a purification.

Identity of a protein, when the analyte is a protein, begins with sequence. UniProt is a public place to check mass, domains and reported processing before you invent a retention time. Community expectations for how protein identifications are reported live with organisations such as HUPO. A chromatogram is not those reports.

Why one molecule comes out later

The stationary phase is a solid, or a coating on a solid, packed in a column. The mobile phase is the liquid you pump or let flow. A molecule that prefers the stationary phase spends longer there and exits later. A molecule that prefers the mobile phase exits earlier. Retention is an average of many brief bindings, not a single click. Changing the mobile phase changes the preference, which is why gradients exist.

Resolution is whether two neighbours become two peaks. It depends on how differently they prefer the phase, on how narrow each band stays, and on how much you loaded. Overloading a column is the everyday way to destroy a separation that worked at analytical scale. Selectivity is the chemical preference. Efficiency is the narrowness of the band. You can have a selective chemistry and still see one blob if the column is damaged, the dead volume is large, or the injection solvent is stronger than the starting mobile phase and the sample smears at the inlet.

Method classes

Reversed phase uses a hydrophobic stationary phase and a polar mobile phase, often water with an organic solvent such as acetonitrile or methanol, plus a modifier. It is the workhorse for many peptides and small molecules. The column silica has a pH window. Outside it, the bed dissolves or the coating leaves. Ion-pairing and buffered systems are variants, not a different science.

Normal phase and hydrophilic-interaction modes keep polar analytes interacting with a polar surface while a more organic mobile phase moves. They are useful when reversed phase retains nothing. They are also thirsty for water control: a little extra water changes retention a lot.

Ion exchange uses a charged group on the resin and a mobile phase of controlled pH and salt. It is a protein and nucleic-acid method as much as a small-ion method. The companion article treats the preparative protein case.

Affinity chromatography uses a ligand that binds one feature of the target: a tag, an antigen, a cofactor. It can look astonishingly clean in one step and still leak the ligand, or still keep every host protein that happens to share that feature.

Size exclusion, or gel filtration, separates by hydrodynamic size when the buffer is chosen so that little true binding occurs. It is a polishing and a buffer-exchange tool. It is a poor first step for a crude lysate because the column has limited volume and every protein is still there, only spread out.

HPLC is the instrument class that pumps those mobile phases at high pressure through fine particles, with an autosampler, a column oven and a detector. Detectors respond to a property, not to a name. Ultraviolet absorbance sees chromophores. A fluorescence detector sees fluorophores. A refractive-index detector sees almost everything and is easily upset by a gradient. A mass spectrometer at the end sees ions and is a second method, not a decorative outlet. Medium-pressure and gravity systems move the same chemistries more slowly, which is often what a folded protein needs.

Expression and lysis reagents upstream of a protein column are a different catalogue class. A supplier list such as NEB products can orient that class. It is not a chromatography method to copy.

A run, and the branch when the trace lies

Equilibrate until the detector baseline is flat in the starting mobile phase. Inject a blank. If the blank has peaks, the solvent, the vial or the injector is dirty, and sample peaks at those times are suspects. Inject a standard that should give a known pattern. If the standard retention has drifted or the peaks have merged, stop. A sample run on a sick system produces confident wrong fractions.

For a real sample, match the injection solvent to the starting conditions. A sample dissolved in strong organic solvent and injected onto a reversed-phase column that starts in mostly water will smear. Clarify particulates. A frit full of debris looks like a broad peak and then like a blocked pump.

If the standard is fine and the sample is a single overloaded hump, inject less or gradient more slowly. If everything sticks, the chemistry is wrong or the pH has flipped the charge you were counting on. If a peak appears in the wash at the end, the gradient never released it; extend the gradient or accept that this mode is holding a contaminant you then have to collect on purpose. Preparative collection should be by fraction and by a test of those fractions, not by a single tube under the tallest line.

A methods record that names column chemistry, particle size, dimensions, mobile phases, gradient, flow and detector is the run. A printed chromatogram without those numbers cannot be repeated.

What the trace supports

Area is not mass unless the detector responds in a known way. Two compounds with different extinction coefficients can have the same ultraviolet area and very different amounts. A protein peak at 280 nanometres is dominated by aromatic residues. A protein with few of them can be abundant and nearly invisible.

Failure modes

Rising pressure is a blocked frit, precipitated salt, or a mobile phase that gelled when an organic solvent met a buffer. Check mixing. Salt crashing in the pump head is a classic when a buffered aqueous line meets a high organic line without a gradient designed to avoid it. Falling pressure is a leak. Retention that shortens over weeks can be a column losing its coating, a mobile phase made with the wrong organic fraction, or a temperature change. Ghost peaks after a dirty sample are carryover. Run a blank after the sample before you believe the next injection. A detector lamp near the end of its life adds noise that looks like minor components. Do not integrate noise because the software can.

Solvents and safety

Organic solvents, acids and column-cleaning agents are chemical hazards. A biosafety cabinet is not the enclosure for a bottle of acetonitrile. Use the chemical control your assessment names, often a ducted hood, and ground and ventilate solvent lines as the instrument manual requires. Waste is mixed solvent and sample. It is not aqueous waste because it looks clear. Protein columns that processed a lysate are also biological waste until inactivated as your rules require. This page is not a purification protocol for a toxin and not medical advice.

Heat, power and water

Retention on many columns moves with temperature. An HPLC oven exists so that "the lab was warmer this week" is not an unrecorded variable. In a building without stable power, a run that dies mid-gradient is a wasted column equilibration and a sample you may not be able to re-inject if the fraction was the whole stock. Do not start a long gradient on a supply that drops every afternoon unless the method can be restarted from a saved sample. Water quality shows up as baseline drift and ghost peaks. Use the water grade the detector requires, and date aqueous buffers so microbial growth in a bottle does not become a peak. Humid air does not replace a capped solvent bottle. Evaporation changes the organic fraction and silently changes retention.

What to send with an enquiry

State the analyte class, analytical or preparative purpose, pressure range, detector, solvent compatibility and fraction collection. Ask which wetted materials and which column chemistries are in scope. Use the scientific instruments catalogue and the quote request. The custom protein expression and purification reference can carry a protein-method question. Read it as a reference. Do not read it as a campaign already underway.