Aptamer Assay

An aptamer assay measures many proteins at once by binding each target with a short, chemically modified single-stranded DNA sequence (an aptamer) instead of an antibody, then quantifying the bound aptamer as a DNA signal. The dominant commercial implementation is SomaLogic’s SomaScan, built on SOMAmer (Slow Off-rate Modified Aptamer) reagents, currently offered at roughly 7,000 and 11,000 plex ¹.

Proteomics splits into three families: affinity assays (aptamer panels like SomaScan, antibody proximity-extension panels like Olink), discovery mass spectrometry (DDA or DIA on a plasma digest), and targeted mass spectrometry (MRM/PRM for a short, absolutely quantified list). Affinity panels win on depth per dollar in plasma because they read low-abundance proteins without depletion or fractionation. Mass spec wins on specificity, because it reads peptide sequence rather than binding affinity.

How it works

Each SOMAmer is a DNA oligo carrying modified nucleotides with hydrophobic side chains (benzyl, naphthyl, tryptamino groups) at the 5-position of dU, which gives the oligo amino-acid-like contact chemistry and lets it fold into a shape with slow off-rate binding to a protein epitope. The assay runs as an equilibrium binding step on streptavidin beads, a polyanionic competitor wash that strips fast-off-rate nonspecific complexes, a photocleavage and second capture, then elution of the aptamers. What gets read is DNA: hybridized to a custom Agilent microarray in older versions, sequenced in newer high-plex configurations. Signal is reported in relative fluorescence units (RFU), which are proportional to protein concentration within a reagent but not comparable across reagents.

Reproducibility is the main reason the platform is used for cohorts. Technical replicate CVs are typically in the single digits to low teens for most reagents, and within-person biological stability across samples drawn a year or more apart is moderate to high for a majority of analytes, which is what makes single-timepoint epidemiology viable ². Aptamer panels have become the workhorse for population-scale pQTL mapping and organ-specific ageing signatures ^{3 4}.

In your own data

The deliverable is an .adat file: a tab-delimited text file with a metadata header block, a row of per-reagent annotations (SeqId like seq.10000.28, Target, UniProt, EntrezGeneSymbol, Dilution, ColCheck), then one row per sample with sample metadata columns followed by RFU columns.

A workable path in R:

library(SomaDataIO)
adat <- read_adat("mystudy.adat")
adat <- adat |>
  dplyr::filter(SampleType == "Sample", RowCheck == "PASS")
anno <- getAnalyteInfo(adat) |> dplyr::filter(ColCheck == "PASS")
x <- log2(as.matrix(adat[, anno$AptName]))

Things to check before anything else:

• Normalization steps applied, listed in the header. Typical chain is hybridization control normalization, then plate-scale, then median signal normalization or ANML (adaptive normalization by maximum likelihood, which pulls your sample toward a healthy reference distribution). ANML is good for population comparisons and bad if your sample is genuinely far from the reference for many analytes.
• Per-sample normalization scale factors. Values outside roughly 0.4–2.5 usually indicate a handling or hemolysis problem.
• Dilution bin per reagent (20%, 0.5%, 0.005% plasma). Reagents sitting in the wrong bin for your analyte’s true concentration saturate or fall to background.
• log2 before anything parametric. Raw RFU distributions are right-skewed.

Common mistakes: treating RFU as ng/mL, comparing RFU between assay versions or vendors without recalibration, and interpreting a single-timepoint outlier as a finding. For a personal profile, the useful unit is your own delta over time on the same platform, not your rank against a reference population.

Limitations

Specificity is per-reagent, not platform-wide. Cross-platform comparisons between aptamer and antibody measurements of the same protein show a broad correlation distribution: some targets agree tightly, a substantial share barely correlate at all, and the two technologies often carry non-overlapping genetic and disease signal ⁵. Treat agreement with an orthogonal method, or a cis-pQTL at the coding gene, as your specificity evidence for a given reagent ³.

Epitope effects are the sharpest trap. A missense variant in the binding site can reduce aptamer affinity without changing protein abundance, producing a strong cis-pQTL and a fake low reading. Orthogonal confirmation by mass spectrometry or immunoassay is the standard remedy, and this is a recognized barrier to clinical use of these panels ⁶.

Privacy matters too: large protein panels carry enough individual-level structure to raise re-identification concerns, so treat an ADAT like genotype data ⁷.

Nothing here is diagnostic. An unusual value on a research-grade aptamer panel is a hypothesis to bring to a clinician, who will order a validated clinical assay if warranted.

Woolf Software builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.