From cells to a gene expression result

From cells to a gene expression result is a chain of biological and analytical choices. Cells make RNA, the laboratory tries to preserve a snapshot of that RNA, and an assay turns the snapshot into a comparison. The comparison is useful when the design, the RNA quality and the analysis agree. It is decorative when any one of those is missing. This page is the pillar for that path. The focused assay is described in RT-qPCR for relative expression. Nothing here is a diagnostic method or a kit insert.

What the result is allowed to mean

Gene expression, in the assays most benches run, means steady-state RNA abundance. It does not automatically mean the rate of transcription, and it does not automatically mean the amount of protein. A higher count can come from more transcription, from slower RNA decay, or from a shift in which cells are present in the tube. Say which of those your design can separate.

The unit matters. A relative RT-qPCR result is a fold comparison under stated assumptions. An RNA-seq result is a count, or a normalised abundance, for a feature in a particular annotation. Neither number is a concentration in the cell unless you built a calibration that earns that unit.

Keeping the snapshot honest

RNA is chemically fragile and biologically short-lived. The interval between the stimulus and the harvest is part of the experiment. So is the way you stop degradation. Cultured cells can go into a lysis buffer that inactivates RNases. Tissues often need to be frozen or placed in a stabiliser immediately, because a warm delay rewrites the transcriptome you meant to measure.

Extraction method is a class choice: organic extraction, silica columns, or magnetic beads, each with a bias for small RNAs or for large ones. Whichever you use, record yield and a purity estimate, and remember that absorbance ratios do not prove the RNA is intact. Integrity is a separate observation. Electrophoretic traces that show ribosomal peaks, or a score computed from those traces, tell you whether the sample is fit for the assay you chose. A badly degraded sample can still produce a number. The number will favour fragments that survived.

Genomic DNA rides along in many RNA preps. It will occupy primers that are not intron-spanning, and it will look like expression. A DNase treatment is a step, not a belief. A no-reverse-transcription control later is the check that the treatment worked.

Annotation comes from a stated genome. Ensembl is one public place to confirm gene models, transcript structure and which exons exist in the species you actually cultured. A primer or a count that assumes the wrong species build will be tidy and wrong.

Two assays, two kinds of claim

RT-qPCR measures the transcripts you already named. You reverse-transcribe the RNA, amplify a short amplicon, and compare quantification cycles between samples after normalising to reference transcripts. It is the right tool when the gene list is short, the material is limited but intact, and you need a comparison you can run locally. It cannot tell you about a gene you did not assay. The minimum information worth recording for a real-time experiment is discussed in the MIQE guidelines. Use that paper as a reporting standard, not as a recipe you paste over your own enzyme.

RNA-seq counts many transcripts at once by sequencing a library made from the RNA. The library type is a scientific choice. A poly(A) selection follows mature polyadenylated RNA and will miss much of the bacterial transcriptome and many non-polyadenylated species. A ribosomal-RNA depletion keeps a broader RNA population and asks for more sequencing and a more careful ribosomal leftover check. Stranded libraries preserve which DNA strand the transcript came from. None of these choices can be repaired in software if you picked the wrong one for the organism.

Sequencing itself is a technology class. A public orientation to one widely used short-read chemistry is the Illumina sequencing overview. Long-read transcript methods answer isoform questions with different library rules. After the reads exist, alignment or pseudo-alignment, quantification, and a differential test are further choices. They belong in the record beside the wet-lab kit names.

Controls that travel with either method

A biological replicate is an independent culture, animal, or donor, processed so that it captures ordinary variation. A technical replicate is the same RNA measured twice. Both can be useful. Only the biological set supports a claim about the condition.

Batch is a quiet covariate. If every treated sample was extracted on Monday and every control on Thursday, the calendar is confounded with the biology. Interleave the conditions. Record the operator, the extraction lot and the day.

Failure that looks like biology

Degraded RNA compresses differences and can create false ones, especially if one group spent longer on the bench. DNA contamination inflates apparent expression for intron-free assays. The wrong annotation builds a confident table of features the organism does not have. A heatmap clusters whatever you feed it, including a batch effect. Look at the sample sheet beside the picture.

Inhibition in RT-qPCR makes a sample look low when it is merely dirty. Dilution or a spike of a known RNA separates those stories. In RNA-seq, a huge ribosomal fraction means the depletion or the poly(A) selection did not do what you paid for. Those reads are a library failure even if the files are large.

Public method collections such as protocols.io are useful for seeing how other laboratories structure a workflow. They are not a substitute for the record of what you did to this batch of cells.

Safety and the limits of the page

Cells, especially human and primate lines, and any sample from an infected source, carry a biosafety decision that your institution makes. RNA extraction chemicals often include chaotropes and solvents that need a chemical risk assessment. This page does not approve a containment level, a diagnostic claim, or the release of expression data as medical information.

Heat, power and moving RNA

In a warm laboratory, "keep on ice" fails if the ice is meltwater and the tubes are floating. Use a cold block and a time limit you actually keep. A freezer that warmed during a power cut is a new experiment: note the temperature excursion before you treat those aliquots as equivalent to undisturbed RNA. Humidity matters when plates and elution tubes sit open. When RNA moves between buildings, name the cold-chain handoff. A box that arrived warm is a reason to repeat the integrity trace.

What to put in an enquiry

A useful sourcing note names the species, the sample type, how the RNA will be stabilised, whether you need polyadenylated RNA or a broader fraction, the number of biological replicates, and the comparison. The nucleic acid analysis pathway is the route for the surrounding sample-to-result work.

The mRNA sequencing reference and the differential expression analysis reference are independent pages you can use to frame a conversation. Ask, through the quote request, whether a quotation is possible. Do not read them as a statement that EVRINTH runs those studies or stocks a particular library kit. State the scientific requirement and ask for the method that would actually be used.