Capillary Electrophoresis and Electropherogram Interpretation

A capillary electrophoresis genetic analyser resolves DNA fragments through a polymer-filled glass capillary, typically 36 cm or 50 cm in length, under a high-voltage electric field. The polymer, POP-4 on the 3130 and POP-7 on the 3500 series (the 3500 also accepts POP-6 for certain fragment-analysis applications), acts as the sieving matrix. Smaller fragments migrate faster and reach the detection window first; larger fragments arrive later. The separation resolution is sufficient to distinguish fragments differing by a single base pair, which is the minimum discriminating unit for STR alleles that share a repeat unit but differ in flanking sequence.

The 3500 and 3500xL instruments carry eight and twenty-four capillaries respectively, running simultaneously in a single electrophoresis run. This parallelism is why a busy forensic laboratory processing case-work can run twenty-four samples with an allelic ladder and size standard in a single injection rather than reinjecting after every eight. The 3130xl runs sixteen capillaries; the legacy 3130 runs four. Indian Central Forensic Science Laboratories (CFSLs) and many State FSLs in India operate a mix of 3130xl and 3500 instruments; the UK Forensic Science Service successor laboratories (Eurofins Forensics UK, LGC Forensics) have largely standardised on 3500xL; the FBI's Combined DNA Index System (CODIS) laboratories in the United States report validation data against 3500xL with GlobalFiler.

Detection is by laser-induced fluorescence. The 3500 series carries a 22 mW, 505 nm solid-state laser. Fluorescently labelled PCR products pass the detection window and emit light at wavelengths characteristic of their dye. The CCD (charge-coupled device) camera captures the emission spectrum. Because multiple dye channels are active simultaneously, a critical pre-run step is spectral calibration to separate the overlapping emission spectra of the different dyes.

Modern STR multiplex kits use five or six fluorescent dyes to label primer sets across the panel. Applied Biosystems chemistry uses the FAM (blue), VIC (green), NED (yellow), PET (red), and LIZ (orange) dye system for five-dye kits such as Identifiler Plus and GlobalFiler. Promega kits (PowerPlex Fusion, PowerPlex 18D) use FL (blue), JOE (green), TMR (yellow), CXR (red), and WEN (orange). Each dye emits over a slightly different emission spectrum, but the spectra overlap significantly.

The spectral matrix (also called the spectral calibration matrix or colour separation matrix) is a correction factor that decouples these overlapping signals. Before a new capillary array is used, or after a routine maintenance event, the laboratory runs a matrix standard that contains all dyes at equal concentration and no DNA. The instrument software (Data Collection on the 3500 series) computes the overlap correction coefficients from that standard run. Every subsequent electropherogram is mathematically corrected by matrix multiplication to yield five separate, idealised dye channels.

If the matrix is outdated, misapplied, or computed on a poorly prepared standard, the correction is imperfect. The result is spectral pull-up (or bleed-through): a tall peak in one dye channel generates a spurious smaller peak in an adjacent channel at the same size position. Pull-up peaks are a major source of false allele calls in inexperienced hands and are the mechanism behind the Adam Scott laboratory contamination case in the UK in 2011, though that case also involved a physical cross-contamination event at sample preparation.

In the United Kingdom, the Forensic Science Regulator's Codes of Practice and Conduct (version 6, 2021) require that matrix correction is validated as part of the laboratory's accreditation package under the Forensic Science Regulator Act 2021. In the United States, SWGDAM's 2017 guidelines require that the matrix be rerun after every capillary replacement or polymer change. In India, FSL laboratories operating under NABL accreditation (ISO/IEC 17025:2017) are expected to follow instrument manufacturer validation protocols, which carry the same requirement.

An allelic ladder is a mixture of commonly observed alleles at each locus in the STR panel, all amplified and pooled at near-equal concentrations. The ladder is co-injected in each run alongside case samples. The analysis software matches each sample peak to the corresponding ladder peak by size, and from that size match assigns the allele designation: 14, 15, 16.2, 17, and so on.

This two-step process (size by internal lane standard, then allele call by ladder comparison) is what makes the system robust to minor run-to-run variation in migration time, which changes slightly with polymer age, temperature, and capillary condition. The internal lane size standard, typically GeneScan 500 LIZ, GeneScan 600 LIZ v2.0, or WEN ILS 500 (Promega), is a set of fragments of known size that is co-electrophoresed in every capillary in every run. The software fits a local size-calling curve from the size standard peaks, which converts migration time to fragment size in base pairs. Allele assignment then matches sample peaks to the ladder's size-calibrated positions.

Alleles that fall outside the ladder range are termed off-ladder (OL) alleles. Off-ladder calls must be verified: the peak may be a true rare allele (a variant allele), a stutter artefact on the low-size side of a genuine allele, or a pull-up artefact from an adjacent channel. Microvariants (e.g. D21S11 allele 29.2, which has three full TCTA repeats and one partial TCTA repeat giving a non-integer allele size) are legitimate off-ladder calls and are reported with their decimal designation. The SWGDAM guidelines and the European DNA Profiling Group (EDNAP) have published population data for the most common microvariants at panels including D21S11, TH01, and FGA.

1. 1. Run internal size standard
   GeneScan 500 LIZ or WEN ILS 500 is co-injected in each capillary. Software fits a local size-calling curve mapping migration time to fragment length in base pairs.
2. 2. Co-inject allelic ladder
   The kit-supplied ladder (e.g. GlobalFiler Allelic Ladder) is run in at least one capillary per plate. Ladder peaks are sized against the same internal standard.
3. 3. Match sample peaks to ladder
   GeneMapper ID-X or OSIRIS assigns an allele number to each sample peak by matching its sized position to the nearest ladder peak within a defined bin window (typically ±0.5 bp).
4. 4. Review off-ladder calls
   Peaks outside bin windows are flagged OL. The examiner checks raw data for microvariant, stutter, or artefact origin before reporting.
5. 5. Apply analytical threshold
   Peaks below the analytical threshold (AT), validated for each kit and instrument combination, are not called as alleles and are not reported.

Stutter peaks are a byproduct of PCR amplification. The DNA polymerase occasionally slips on the repeat unit during extension, producing a product one repeat unit shorter (n-4 stutter at a TPOX tetranucleotide locus, n-3 at a trinucleotide) or, less commonly, one unit longer (plus-stutter). Stutter peaks consistently appear immediately below (and occasionally above) each genuine allele peak in the electropherogram.

Stutter ratios, the ratio of the stutter peak height to the parent allele peak height, are locus-specific and have been characterised for every locus in every major STR kit. Typical stutter ratios are in the range of 5 to 15 percent for most CODIS core loci. A peak below the stutter ratio threshold for that locus and that allele size is treated as a stutter artefact rather than a genuine allele. SWGDAM's 2017 guidelines and the ENFSI DNA Working Group's 2016 document both require that stutter thresholds be validated empirically during kit and platform validation, not taken from the manufacturer's documentation alone.

The interpretive complication arises in mixture samples: if a genuine allele from a minor contributor happens to fall at the stutter position of a major contributor's allele at the same locus, the minor contributor's allele may be masked or misclassified. This is the central problem that probabilistic genotyping software was designed to address, and is covered in depth in mixture deconvolution and probabilistic genotyping.

In the UK, R v. Adams (No 2) (1998) involved, among other issues, the question of how statistical weight was assigned to DNA evidence when the raw data required interpretive judgment. The court was not dealing with stutter per se, but the case established the principle that the examiner's interpretive decisions are part of the evidence and must be disclosed. In Australia, the Forensic Science Regulator equivalent, NATA (National Association of Testing Authorities), audits stutter threshold validation records as a standard accreditation item.

Drop-out is the failure of an allele to amplify to a detectable level. At low template amounts, the Poisson statistics of PCR starting material mean that a given allele may be present in only one or two initial copies. If those copies fail to enter the reaction at the denaturation step, that allele produces no signal above the analytical threshold. The result is an apparent homozygote at a locus where the true genotype is heterozygous, or a locus where only one of two genuine alleles is reported.

Drop-in is the complement: the sporadic amplification of a low-level contaminant allele that was not present in the original evidence. Drop-in peaks are characteristically small (typically below 200 relative fluorescence units, RFU, in most validated workflows), appear at only one or two loci across the profile, and do not recur in re-amplifications. The distinction between a minor contributor's genuine allele and a drop-in peak is a practical challenge in every high-sensitivity forensic workflow.

The stochastic threshold (ST) is the minimum peak height above which an allele is considered reproducibly amplified and therefore interpretable. Peaks between the analytical threshold and the stochastic threshold are visible but uncertain: they may represent genuine alleles that are dropping towards zero, or they may be artefacts. The SWGDAM guidelines require that the ST be empirically determined from the laboratory's own data and documented in the laboratory's validated interpretation guidelines. UK laboratories working under the Forensic Science Regulator's codes apply a similar concept under the term probabilistic threshold, and Australian NATA-accredited labs document it as part of the validation package for each kit.

Every interpretation decision the examiner makes on the electropherogram must be documented. This includes: which peaks were called as alleles and why, which peaks were treated as stutter or artefact and why, which loci were treated as potential drop-out loci, and whether the stochastic threshold was triggered. In accredited laboratories, this documentation typically lives in a case notes file that is retained alongside the raw data files (FSA files on ABI instruments) and is subject to discovery in any criminal or civil proceeding.

In the United States, the FBI's Quality Assurance Standards (QAS) for Forensic DNA Testing Laboratories require that raw data be retained in a recoverable format and that the case file document all interpretive decisions. CODIS upload requires a declaration that the profile was generated by a validated and QAS-compliant process. The allele calls from the electropherogram feed directly into the random match probability and likelihood ratio calculation reported to the court. In England and Wales, the Criminal Procedure and Investigations Act 1996 (CPIA) requires disclosure of all material that could assist the defence, including the raw electropherogram data and any notes on artefacts encountered. The 2021 Forensic Science Regulator Act put that disclosure obligation on a statutory footing for the first time.

In India, the Bharatiya Sakshya Adhiniyam 2023 (BSA § 39, replacing IEA § 45) governs the admissibility of expert opinion, and forensic DNA testimony routinely involves the production of printed electropherograms and instrument output files as supporting exhibits. Indian FSL laboratories operating under CBI or CFSL authority maintain case files that include the raw electropherogram printouts. The challenge, documented in Lok Sabha committee reports on the DNA Technology Bill, is that chain-of-custody documentation for instrument outputs is not yet uniformly standardised across State FSLs, a gap the proposed DNA profiling rules would address.