Every human's DNA is 99.9% identical — but at STR (short tandem repeat) loci, short sequences like GATA repeat back-to-back a variable number of times, and that repeat count differs between people. Each person carries two alleles per locus (one from each parent), so a locus is described by two numbers, e.g. "14, 17 repeats".
PCR makes billions of copies of just those loci so there's enough DNA to measure: each cycle roughly doubles the copy count.
copies(n) = copies(0) · 2ⁿ
size(bp) = base(bp) + repeats · unit(bp)
Capillary electrophoresis then pulls the amplified fragments through a gel-filled capillary under an electric field. Shorter fragments (fewer repeats) slip through faster and reach the laser detector sooner; longer fragments lag behind. The detector logs each arrival as a peak on the electropherogram, plotted by fragment size — that's the DNA profile.
- Locus — pick which STR marker to inspect; each has its own typical repeat-count range and its own color in the graph.
- Suspect — four candidate profiles are generated per case; only one (if any) shares every allele with the crime-scene sample.
- Matching rule — a locus "matches" only if both alleles are identical between the two profiles. A single mismatched locus is enough to exclude a suspect, exactly as in real casework.
Real-world relevance: this PCR → capillary electrophoresis → peak-comparison pipeline (CODIS-style, ~13–20 core STR loci in practice) is the actual workflow forensic labs run to identify a contributor or exclude a suspect from crime-scene DNA.