Plasma and the dynamic range problem

Plasma and serum are hostile mixtures for a mass spectrometer because a few proteins are enormously more abundant than the rest. The dynamic range problem is that physical fact. A digest of neat plasma mostly becomes peptides of albumin, immunoglobulins and their immediate neighbours, and the instrument reports them confidently. Bottom-up proteomics in plain language describes how any complex mixture hides its minor components. Plasma is the mixture where that hiding is the whole experimental design.

What occupies the mass

Albumin alone is a huge share of plasma protein mass. Immunoglobulins are the next familiar block. A small set of other classical plasma proteins rides with them. Together they are a huge fraction of what a pipette picks up. The proteins people often hope to see, including many signalling proteins and leakage markers from tissue, sit far down a range that is commonly described as spanning many orders of magnitude. This page will not invent a fold number for that span. "Many orders" is enough to explain why a chromatogram is a few towers and a low baseline.

The peptides from the towers ionise well, fragment well, and fill the data-dependent picker. The false discovery rate on the resulting list can look excellent, because the matches are genuinely correct. They are correct about the abundant fraction. A tidy error rate is not evidence that the low-abundance fraction was measured. PeptideAtlas builds of plasma show how much repeated observation still clusters on proteins the community can see with ordinary depth. Read that as a map of difficulty, not as a menu of markers.

Serum and plasma are not the same sample. Serum is what remains after clotting, and the clot removes fibrinogen and changes other proteins. Plasma is collected with an anticoagulant. EDTA, citrate and heparin are different chemical backgrounds. Heparin in particular is unkind to some mass spectrometry workflows. The anticoagulant belongs in the sample record next to the species and the time to freeze. Haemolysis dumps red-cell proteins into the same dynamic range problem and paints the low end with a new abundant set. A pink sample is a different experiment.

What a plain digest can honestly return

A plain digest, desalted and injected at a load the column can bear, is a good way to see the abundant plasma proteome. That can be the right question. Albumin modifications, immunoglobulin composition, and other high-mass proteins are legitimate research targets. Pretending the same injection also surveyed the trace proteome is how reports over-claim. The chromatogram will tell you: a few very large peaks, a busy but shallow middle, and a baseline that is not a catalogue of thousands of equal proteins.

Quantification among those abundant proteins has its own traps. Their signals sit near saturation. A label-free comparison of albumin between two samples can be a comparison of overloaded peaks. Dilution, or a method aimed at that protein, is the honest tool when the protein is a major component. The dynamic range problem is not only "we cannot see the small ones". It is also "the large ones can be too large to measure fairly on the same injection".

Depletion, and what leaves with the target

Depletion columns use antibodies, or other binders, against a panel of abundant plasma proteins. The flow-through is what you digest. The bound fraction is enriched in albumin and its panel partners, and you usually discard it or keep it for a different question. Two consequences follow.

First, the mass spectrometer can now spend time on proteins that were previously buried. Identifications move down the abundance scale. They do not reach the bottom of it. A depleted sample still has a dynamic range, just a shorter one, and a short data-dependent run of the flow-through is still a slice.

Second, proteins that travel bound to albumin or to immunoglobulins are depleted along with their carriers. The literature calls this the hitchhiker problem. A protein that vanishes after depletion may have been removed because it was a passenger, not because it was never in the plasma. If your question might involve carriers and cargo, analyse the bound fraction or run a neat sample in parallel. Do not interpret absence in the flow-through as biological absence.

Depletion columns also have a capacity. Overloading them leaks albumin back into the flow-through and restores the original problem in disguise. The manufacturer's load range is the instruction. Accurate micropipetting technique matters here because a "small volume of plasma" that is actually a large one will saturate the resin. Record the volume and the column's stated capacity class.

Enrichment classes that prefer a different slice

Nanoparticle and related enrichment classes work by forming a protein corona, or by some other selective binding, so that a subset of proteins is collected onto a surface and the bulk albumin is partly left behind. Other methods fractionate by chemistry before digestion. Each class shows a different slice. Each class also loses proteins that do not bind. Comparing a nanoparticle eluate with a depleted flow-through as if both were "plasma" will invent differences that are the methods.

None of these preparations turn one short data-dependent acquisition into a complete plasma proteome. Depth still depends on gradient length, fractionation, and how much of the remaining range the instrument is allowed to sample. A claim of completeness from one brief run is not credible, depleted or not. Say which slice you measured, and under which false discovery rate. HUPO plasma-related initiatives are a public place to see how cautious that language is supposed to be. They are not a certificate for a local sample.

Neat, depleted, and enriched slices

PRIDE deposits of plasma studies are worth reading for whether the method section names the slice. NIST reference-material thinking is a useful attitude: know what the sample is supposed to represent before you compare two numbers. A local plasma aliquot is not a reference material unless it was made as one.

Failure modes that belong to the range

Leakage through an overloaded depletion column looks like "depletion did nothing". Check capacity before you conclude the biology is all albumin.

A huge apparent change in a trace protein, seen in only one undepleted run, is often a sampling accident among the tall peaks. Replicate it in a preparation that can actually see that protein.

Batch effects between depletion columns are real. A new lot of resin can change the flow-through. Bridge lots with a shared quality-control plasma if the study is long. That is study design, not a complaint about dynamic range in the abstract.

Mixing serum and plasma in one table mixes clotting chemistry with biology. Separate them in the design.

Research chemistry, not a diagnostic assay

Everything above is research sample chemistry. It helps a laboratory decide which slice of plasma to measure and how to describe it. It does not validate a test, set a clinical cut-off, or interpret a person's result. Biosafety and human-sample governance are institutional decisions: consent, identifiers, and containment sit with the committee that approved the work. A mass list does not replace that approval. Do not write a diagnostic sentence because the dynamic range was "addressed".

Cold-chain handoff

Plasma that thaws on the way from a collection site to the analytical laboratory is a changed sample. Proteases work during a partial thaw, proteins can precipitate, and a second freeze does not rewind the chemistry. The handoff specification should say who freezes, at what temperature class, how the cold is maintained in transit, and what a receiving laboratory does if the ice is gone or the tube is warm. In a long journey through heat, plan the coolant for the duration you actually face, and record the condition on arrival. A tube that arrived cold and a tube that arrived thawed should not share a statistical group. HUPO sample-collection discussions are background for why that sentence belongs in a protocol. Your courier plan is local.

What to say in an enquiry

State whether the biological question is about the abundant proteins or about a lower-abundance fraction. Name plasma or serum, the anticoagulant if it is plasma, the species, the volume class, and whether depletion or another enrichment is already assumed. Say if hitchhiker proteins would matter. Say how many samples, and whether a quantitative comparison is required.

Those facts can be discussed against the shotgun discovery proteomics reference, the protein identification by LC-MS/MS reference and the differential abundance reference. Put them in the quote request. A method discussion should name the slice and refuse a completeness claim the design cannot support. That refusal is the useful part of the conversation.