The MIQE idea in everyday language

A fold change is easy to paste into a slide and hard to audit. Quantitative PCR reports a cycle at which fluorescence crossed a threshold. That cycle, usually written Cq, becomes a comparison only after someone states the sample, the chemistry and the arithmetic. The MIQE guidelines, from Bustin and colleagues in Clinical Chemistry, exist so a second scientist can make that judgement. The paper is at doi:10.1373/clinchem.2008.112797. This page puts that idea into everyday language for a research bench. The wider path from cells to a result sits in from cells to a gene expression result. The assay mechanics of a relative comparison sit in RT-qPCR for relative expression.

Who the note is for

MIQE serves anyone who will publish, archive or hand over a quantitative PCR result: the person who ran the plate, the colleague who will repeat it next year, and the reader who must decide whether the fold change can carry the sentence in the abstract. The decision it supports is narrow. Either the record is complete enough to judge the number, or the claim has to shrink until it matches the record. It is a research reporting habit. It is not a clinical authorisation and not a kit insert.

If the biological question has already outgrown a handful of named transcripts, the same honesty applies to a sequencing design. The mRNA sequencing enquiry reference and the differential expression analysis enquiry reference are places to frame that conversation. They are enquiry references. A method can be discussed. The quote request is where the scientific requirement goes.

What the fluorescence number is made of

A real-time instrument watches a fluorescent signal grow as a polymerase copies a short amplicon. Early cycles sit in the noise. At some cycle the signal clears a threshold the software, or the user, has set. That cycle is the Cq. Older papers say Ct or Cp for the same family of ideas. MIQE prefers Cq and asks you to say how it was called, because a threshold set by eye and a threshold set by a software model can move the number.

Cq is related to how much target entered the reaction, and also to how efficiently each cycle doubled, to whether inhibitors slowed the enzyme, and to whether the reverse transcription that built the cDNA was comparable across samples. Two tubes with the same Cq can still differ in meaning if one was inhibited or if one reference transcript moved with the treatment. The reporting idea is to put those dependencies on the page so the Cq is interpretable.

Specificity sits underneath the arithmetic. If the primers amplify a second product, or genomic DNA, or a primer dimer, the fluorescence is no longer a clean proxy for the transcript you named. NCBI Primer-BLAST is a public place to see what else a pair might bind. A melt curve or a gel is the local check that the product in your buffer is the one you designed.

Information classes, without a copied volume table

Think of the report as seven classes of fact. None of them is a microlitre recipe. The enzyme you used has its own instructions, and those instructions stay with the tube.

The sample class names the organism, the tissue or line, the contrast, the time point, and how the material was stabilised. RNA handling belongs here. Heat, a delay on the bench, and a thaw all change what the later assay can see. The practical side of that handling is in protecting RNA during extraction.

The extraction class names the method family, whether DNase was used, and a quality observation: yield, an absorbance ratio if you measured one, and an integrity trace if the assay depends on intact RNA. Absorbance ratios do not prove integrity. Say which observation you have.

The reverse-transcription class is easy to skip because it happens before the plate that everyone remembers. One-step mixes combine reverse transcription and PCR. Two-step workflows make cDNA first. Priming can be oligo-dT, random, gene-specific, or a blend. Those choices change which transcripts are copied. Record the class, the RNA input class, and the enzyme class.

The oligonucleotide class identifies the target: gene symbol, accession, where the amplicon sits, intron spanning or not, and the primer sequences or a stable public identifier. A paper that says "primers as previously described" and points at a broken link has not identified the assay.

The validation class covers specificity, a standard-curve efficiency, the range over which the response was linear, and the no-template control. Efficiency and reference-gene choice are developed in standard curves and primer efficiency and choosing reference genes instead of hoping. MIQE asks for justification of the reference genes in the experiment you ran.

The analysis class names the normalisation, the calibrator, the software and version, how outliers were handled, and the statistical test. Technical replicates and biological replicates are both worth counting, and they answer different questions.

Reagent brands and lot identifiers sit in the record when you have them. They help a repeat. They are not a substitute for the efficiency and the controls.

When a field is missing, shrink the claim

Start from the sentence you want to write. If the sentence is "transcript A was about twice as abundant in treated cells as in the calibrator, relative to two reference genes that were stable in this set", every clause needs a line in the notebook. Walk the classes above and mark gaps.

A missing no-template control means you cannot show that the signal is independent of contamination and primer artefact. Stop and rerun with that control, or drop specificity language from the claim. A missing no-reverse-transcription control, on an assay that can see genomic DNA, means apparent expression may be DNA. That branch goes back to DNase and primer design, not to a stronger statistical test.

A missing efficiency means you may still describe Cq ranks cautiously, and you should not dress the result in a precise fold change that assumes every cycle doubled. A single untested reference gene means the comparison is provisional until stability is checked in this sample set. If the stability check fails, choose other candidates or change the normalisation. Do not keep the gene because a different tissue paper used it.

If several classes are missing and the samples cannot be rerun, the honest product is a methods gap list, not a biological conclusion. Public method notes on protocols.io can show how other groups structure a workflow. They do not fill in the facts of your plate.

Quiet ways a tidy fold change misleads

Contamination that hits every well, including the unknown samples and not the empty notation in the notebook, shifts all Cq values and can still look internally consistent. The no-template well is the witness. Primer dimers do the same job in late cycles, especially when the true target is scarce. A melt peak at the dimer position, or a small band, sends you back to primer design or annealing, which is an optimisation of this pair in this buffer.

Unequal RNA integrity between groups imitates differential expression. If treated tumours sat longer in a warm clinic before freezing, and controls were banked immediately, the Cq pattern may be degradation. Write the delay down. Inhibition from residual salt, phenol or haem makes a sample look low. A dilution series of that sample, or a spike of an exogenous RNA, separates "less target" from "the enzyme struggled".

Software defaults are part of the method. Baseline windows and threshold rules should be the same across the plates you compare, or the difference should be declared. Excluding wells after you have seen the biological answer needs a rule you wrote beforehand.

Research use, and whose rules apply

This page does not approve a diagnostic claim, a biosafety level, or the release of a Cq as medical information. Human and animal specimens, and any material from an infectious source, stay under the institutional rules that already govern your laboratory. Extraction chemistries can include chaotropes and solvents. Follow the hazard note for the kit class you use. A transparent research report and a regulated assay are different documents.

Heat, humidity, and a freezer that blinked

In a warm laboratory the interval between harvest and lysis is part of the MIQE sample record, because RNase does not wait for the write-up. Humidity matters when plates are sealed late or elution tubes sit open. A power cut that warmed a freezer is a storage event: note the excursion before those aliquots are treated as equivalent to undisturbed RNA. The reporting habit is the same one MIQE asks for. The local climate simply makes the sample line harder to ignore.

What to put on an enquiry

For reagents, name the assay type, the organism, the amplicon scale, and whether you need a real-time mix, a reverse transcriptase class, or plasticware matched to an optical block. Catalogue families live under the molecular biology catalogue. For a study that needs many transcripts, say so, and point at the nucleic acid analysis pathway for the surrounding path from sample to result.

Use the sequencing and differential-expression pages as enquiry references when you want those methods discussed. State species, RNA condition, library preference, replicates and the contrast on the quote request. Ask whether a quotation is possible. The pages describe methods a laboratory can ask about. They are not a statement that a particular study is already booked.