A human genome array is a collection of probes designed to recognize selected nucleic-acid sequences. The exact platform determines the question it can answer. Start with its product name and assay type, then identify the sample and reported quantity. A gene-expression signal, a genotype call and a copy-number estimate carry different meanings.
This guide helps students and researchers read array descriptions and public datasets. A clinical result requires interpretation through the laboratory and clinician responsible for that test.
Separate the three common uses
| Array type | Main input and output | Interpretation boundary |
|---|---|---|
| Gene expression | RNA-derived labeled material; relative transcript-associated signal. | Expression depends on tissue, collection and processing; it does not directly measure protein activity. |
| SNP genotyping | DNA; calls at the variants represented by the assay. | Coverage is defined by the probe design. Variants outside that design can be missed. |
| Copy-number analysis | DNA; estimates of gains or losses across genomic regions. | Resolution, mosaicism and detectable event types depend on the platform and analysis. |
A platform may combine functions, but its documentation must establish that capability. Check the intended use as well as the name: a research array and a clinically validated test have different evidentiary and reporting contexts. NHGRI's microarray glossary explains the basic probe-binding principle.
Where the U133 Plus 2.0 fits
The GeneChip Human Genome U133 Plus 2.0 is a gene-expression array. Its manufacturer information sheet describes probes used to measure transcription and identifies the product as research-use only. It should be interpreted as an expression platform, rather than a universal assay for every human genomic variant.
The design draws on early genome and transcript databases, including 2001 and 2003 sequence resources. Historical counts of genes or transcripts reflect those annotations. When reusing a dataset, map the actual probe-set identifiers to an appropriate, documented annotation release. Several probe sets may map to one gene; a probe set may have ambiguous or changed mapping. Keep the original identifiers alongside any modern gene symbols.
Audit a public dataset before comparing groups
- Identify the platform. Record the GEO Series accession, platform accession, tissue, organism and assay.
- Build the sample sheet. Record independent subjects, group assignments, paired samples, collection times and available technical variables.
- Find the data level. Determine whether the download contains raw files, normalized intensities or an already filtered result table.
- Inspect distributions and outliers. Check that samples are comparable and look for processing batches associated with biological groups.
- Specify the comparison. Define the numerator and denominator, relevant covariates and how repeated or paired measurements will be modeled.
- Preserve the analysis record. Save software versions, transformation, normalization, annotation, exclusions and statistical settings.
NCBI's GEO2R documentation explains that its microarray analysis uses submitter-supplied processed tables and warns that heterogeneous processing and study quality affect interpretation. It is a useful exploration tool when its model fits the study. A complicated paired or confounded design deserves an explicitly specified analysis.
Interpret log changes and batch effects
Review effect size together with uncertainty and multiple-testing adjustment. A small adjusted P value alone tells little about practical importance. Inspect the probe mapping and expression range before turning a statistical hit into a biological claim.
Choose arrays or sequencing for the actual task
Arrays offer measurements tied to a predefined probe design and can be useful for established assays and compatible historical datasets. Sequencing can address questions outside a fixed probe set, with its own library, depth, alignment and analysis requirements. Compare the exact target, sensitivity needs, specimen quality, reference data, validated workflow and complete analysis cost.
A sensible first question for a service provider is: “What result would distinguish my biological alternatives, and what features would this assay miss?” Request an example report before commissioning a study. For a clinical question, discuss the appropriate validated test with the treating team; more data alone does not ensure a more useful diagnosis.