What Does an Omics Assay Measure?
A molecular assay does not measure “the biology” in general. It interacts with a particular molecule or molecular property, produces an analytical signal, and records that signal in a specified unit. To interpret an omics result, keep those pieces separate: the biological target, the analytical observation, and the recorded unit. Confusing them can make a technically accurate measurement sound like evidence for a different biological claim.
For example, an assay may target RNA molecules but record sequencing reads associated with those molecules. The reads are evidence about RNA abundance; they are not themselves RNA molecules. Similarly, a mass-spectrometry assay can produce a signal associated with a metabolite, but signal intensity is not automatically a concentration. What is recorded depends on the assay and its measurement process.
A useful way to interpret a result is to ask, in order: What material was assayed? What property or molecule was targeted? What did the instrument or analysis directly record? What unit describes that record? Finally, what biological interpretation does the measurement support, and what does it not establish by itself?
Five Molecular Layers and Their Targets
The five major layers in this course differ first in the biological material or property under study. Their typical recorded units also differ, although the exact output depends on the assay design, platform, preparation, and processing. The examples below are common forms, not rules that every assay must follow.
| Layer | Biological target | Examples of recorded observations and units | Interpretive caution |
|---|---|---|---|
| Genomic | DNA sequence or other properties of genomic DNA | Sequence-read evidence; a call or score for a DNA feature | DNA provides information about sequence or DNA state; it does not directly measure gene activity. |
| Transcriptomic | RNA molecules, often as evidence about transcription | Read or fragment counts; estimated abundance in method-specific units | RNA abundance is not identical to protein abundance or biological function. |
| Epigenomic | Molecular properties of chromatin, including DNA methylation or accessibility | Fractions, enrichment values, or other assay-specific signals | A signal at a genomic region is not, by itself, proof of a downstream expression change. |
| Proteomic | Proteins, often using evidence from their constituent peptides | Peptide or protein signal intensity; estimated amount or concentration when calibrated | Detected peptide evidence and estimated protein abundance are related but not interchangeable. |
| Metabolomic | Small molecules involved in metabolism and related processes | Feature intensity; estimated amount or concentration for identified compounds | An observed feature may not have a confirmed compound identity or calibrated concentration. |
The table’s “recorded observations” column describes common output types, not universal units. A concentration might be reported in mol/L or a submultiple such as μmol/L. By contrast, an instrument intensity may be reported in arbitrary units, which are meaningful within a defined measurement setup but are not a physical concentration. A count is a number of recorded events under a specified counting procedure. These units answer different questions and should not be compared as though they shared a common scale.
The sample is also distinct from the target. A sample might be a tissue specimen, a blood draw, or a cultured cell preparation. The target is the molecule or property examined within it. The assay’s recorded observation is then attached to that sample and to a particular molecular target or feature. Naming the sample alone does not reveal what was measured.
Worked Examples: Matching Targets to Measurements
Worked Example: DNA Evidence Is Not an Expression Measurement
A laboratory receives a hypothetical tissue specimen labelled Sample A. A DNA-based assay produces sequence evidence at a selected genomic region. The result is recorded as 18 supporting reads for one observed sequence and 22 for another. What layer is this, and what can the record support?
The assayed material is DNA, and the target is a property of genomic DNA at the selected region. This is a genomic assay. Its direct recorded observations here are read counts: 18 reads and 22 reads. The values are counts of sequence evidence under this assay, not measurements of RNA, protein, or gene activity.
The arithmetic total is \(18+22=40\) reads at the region in this example. That total describes the stated read evidence; it is not automatically the number of DNA molecules in the specimen. The assay’s sampling and processing determine how many reads are generated. Nor does the larger count alone establish that the corresponding DNA sequence is more active: the assay targets DNA sequence evidence, not transcription.
The interpretation is therefore limited but useful: Sample A yielded the stated DNA read evidence at the selected region. A claim about RNA expression would require an RNA-targeting measurement, and a claim about protein abundance would require protein-targeting evidence.
Worked Example: Choosing Between RNA and Protein Measurements
A hypothetical cell experiment asks two questions about the same treatment: did the treatment change the amount of a particular RNA, and did it change the concentration of the corresponding protein? One assay records 1,240 RNA-associated fragments in the treated sample and 980 in the comparison sample. A separate calibrated protein assay reports 6.2 ng/mL and 5.0 ng/mL, respectively. Match each question to its measurement and interpret the units.
The RNA question calls for the transcriptomic assay. Its stated observations are 1,240 and 980 RNA-associated fragments. The treated sample has \(1{,}240-980=260\) more recorded fragments than the comparison sample. This is a difference in fragment counts, not a difference of 260 RNA molecules; the count depends on the assay and the way fragments are generated and recorded.
The protein question calls for the proteomic assay. Its measurements are concentrations, both in ng/mL. The difference is \(6.2-5.0=1.2\) ng/mL, so the treated sample’s reported protein concentration is 1.2 ng/mL higher in this hypothetical comparison. Because the protein assay is described as calibrated and reports concentration, its unit has a concentration interpretation. The RNA fragment count does not have that same unit.
These observations address related but distinct questions. A higher RNA-associated count does not guarantee a higher protein concentration, and these two sample measurements alone do not establish that the treatment caused either difference.
Worked Example: Separating an Epigenomic Fraction from a Metabolite Signal
A hypothetical analysis of Sample B produces two results. At a selected DNA region, an epigenomic assay reports a methylation fraction of 0.42. A separate mass-spectrometry assay detects a feature with intensity 730 arbitrary units; the feature has not been confidently identified as a particular compound. Identify the targets, the recorded units, and the appropriate interpretation.
The first result is epigenomic: the target is a DNA methylation property at the selected region, and the recorded value is a fraction. Expressed as a percentage, it is \(0.42\times100=42\%\). The percentage is a different expression of the same fraction, not another measurement. This value describes the assay’s estimated methylation fraction for that region; it does not by itself establish a change in gene expression.
The second result is metabolomic in its intended molecular layer because the assay is being used to examine small-molecule features. Its direct observation is an intensity of 730 arbitrary units. Since the feature is not confidently identified, the result should be described as an observed signal rather than as the concentration of a named metabolite. The value cannot be converted to a concentration from the information given.
The two numbers have different meanings and units. It would be incorrect to compare 0.42 directly with 730 or to describe either as a general measure of “molecular activity.” Their interpretation depends on the target and the measurement scale.
A Short Justification: Why Signal and Target Must Be Separated
The distinction between target and recorded unit follows from how an assay works. The assay interacts with material in the sample; an instrument or analytical procedure produces observations; a measurement process links those observations to a biological target. That link may be direct in some senses and indirect in others. For example, a count of sequence reads is direct as a count of reads but indirect as a measure of the amount of starting RNA or DNA.
Determine whether the assay is applied to DNA, RNA, chromatin, proteins, or small molecules in the sample.
Specify what feature of that material the assay is designed to examine, rather than using a broad label such as “gene activity.”
Decide whether the output is a count, fraction, intensity, score, amount, or concentration. Do not substitute one kind of unit for another without a valid calibration or transformation.
State what the assay records and only then describe the biological interpretation supported by that record.
This is not merely a naming convention. If the direct observation is an intensity, calling it a concentration assumes an additional relationship between intensity and amount. That relationship generally requires suitable calibration and assay-specific assumptions. If those are absent, the intensity remains an intensity. Likewise, observing a property of DNA does not directly measure the RNA or protein that may be related to it.
Diagnostics, Comparisons, and Common Pitfalls
Before interpreting an omics result, check that the assay’s intended target matches the question. A question about DNA sequence requires DNA-based evidence; a question about RNA abundance calls for RNA-based evidence; and a question about a protein or metabolite calls for a corresponding molecular assay. Questions about chromatin properties require an epigenomic measurement. An assay can be technically successful while still being a poor match for the biological question.
Next, inspect whether the reported unit supports the wording of the conclusion. A count supports a statement about recorded counts. A fraction supports a statement about a proportion or fraction. An intensity supports a statement about signal intensity. A calibrated amount or concentration supports a statement in that physical unit, subject to the calibration and assay assumptions. If a result is reported in arbitrary units, do not silently treat it as an absolute amount.
A further diagnostic is to ask how directly the assay identifies the target. Some methods measure a signal associated with a target rather than observing the target molecule in isolation. Peptide evidence can inform a protein measurement; a mass-spectrometry feature can be consistent with a small molecule; a sequence read can provide evidence about DNA or RNA. The strength of an identification and the unit of the reported value should be kept separate. A strong signal does not necessarily guarantee an unambiguous identity.
The five layers are complementary, not interchangeable alternatives on one common scale. Genomic measurements characterize DNA; transcriptomic measurements characterize RNA; epigenomic measurements characterize chromatin-related properties; proteomic measurements characterize proteins; and metabolomic measurements characterize small molecules. Comparing them can help connect different biological levels, but a difference in one layer does not automatically imply a corresponding difference in another. Their units and measurement processes need not align.
Common mistake: “The assay reported a large signal, so the sample must contain a large amount of the named molecule.” The reasoning skips two checks: whether the feature has been confidently identified and whether the signal has been calibrated to amount. The correction is to report the observation at its actual level—for example, “a feature intensity of 730 arbitrary units”—and make a concentration claim only when identification and calibration support it.
- Mislabeling the layer: A result about DNA is not an RNA result simply because the DNA feature is related to a gene.
- Equating related molecules: RNA and protein measurements concern distinct molecular targets; one cannot be substituted for the other.
- Dropping the unit: “The value increased” is incomplete when the difference is in counts, concentration, or an assay-specific signal.
- Overstating an indirect result: A molecular signal can support an interpretation without measuring the downstream biological process directly.
Check Your Understanding
For each question, identify the appropriate molecular layer and state what the unit does—and does not—tell you.
- A laboratory wants to characterize DNA sequence in a tissue sample. Which layer is relevant, and would a DNA assay directly measure protein concentration?
- An RNA-targeting assay records 450 fragments for one sample and 500 for another. What is the observed difference, and what is its unit?
- A chromatin assay reports a fraction of 0.30 for a selected property. Express this as a percentage and name the relevant layer.
- A protein-related signal is reported as 900 arbitrary units, with no calibration information. Can it be reported as 900 ng/mL? Why or why not?
- A mass-spectrometry feature has not been confidently identified. What is the most defensible way to describe its recorded intensity?
Solutions. (1) The relevant layer is genomic because the target is DNA sequence. A DNA assay does not directly measure protein concentration; that requires protein-targeting evidence and an appropriate concentration measurement. (2) The observed difference is \(500-450=50\) fragments. The unit is fragments, not RNA molecules or a concentration. (3) The layer is epigenomic if the property is chromatin-related. The percentage is \(0.30\times100=30\%\). (4) No. Arbitrary units describe an assay signal; without appropriate calibration, the result cannot be stated in ng/mL. (5) Report an observed feature intensity in the stated arbitrary units, and note that the compound identity is uncertain rather than naming a metabolite as confirmed.