Sanger Sequencing Chemistry and Modern Methods

A normal deoxyribonucleotide triphosphate (dNTP: dATP, dCTP, dGTP, dTTP) carries a 3'-hydroxyl (-OH) group on the deoxyribose sugar. DNA polymerase extends a growing chain by attacking this 3'-OH with the alpha-phosphate of the incoming dNTP, releasing pyrophosphate and forming a 3'-5' phosphodiester bond. Chain elongation continues as long as a 3'-OH is available.

A 2',3'-dideoxynucleotide triphosphate (ddNTP: ddATP, ddCTP, ddGTP, ddTTP) is identical to its deoxy counterpart except that the 3'-hydroxyl is replaced by a hydrogen atom (3'-H). When a DNA polymerase incorporates a ddNTP, it forms a normal phosphodiester bond with the previous nucleotide's 3'-OH, but the newly incorporated ddNTP has no 3'-OH. The chain cannot be extended further. Synthesis stops at that position.

In the original Sanger protocol (1977), four separate reactions were run, each containing a template, a primer, all four dNTPs, a small amount of one ddNTP, and DNA polymerase (originally the Klenow fragment of E. coli DNA Pol I, later Sequenase, a modified T7 DNA polymerase). In the ddATP tube, synthesis terminates wherever an adenine would be incorporated; some fraction of the time, the ddATP is incorporated instead of dATP, so termination products of different lengths accumulate at every A position in the template. The four tubes together produce four nested ladders of fragments, each ending at a different base type. Separating all four lanes on a denaturing polyacrylamide gel and reading up the autoradiogram from shortest to longest gave the sequence directly: whichever lane had the shortest unique band corresponded to the first base after the primer.

The key insight is probabilistic: the ratio of ddNTP to dNTP in each reaction is tuned so that incorporation of the terminator occurs with low frequency at each A (or C, G, or T) position, generating a population of molecules that terminates at every single instance of that base across the read length. Too high a ratio of ddNTP and all chains terminate at the first occurrence; too low, and most chains run to full length without termination.

Later refinements used thermostable DNA polymerases (AmpliTaq FS, then Taq FS Gold) in a cycle sequencing format: repeated denaturation-annealing-extension cycles generate exponentially more termination products from a linear template, reducing the amount of DNA needed and enabling sequencing from plasmids, PCR products, or even small amounts of environmental DNA.

The transition from radioactive to fluorescent detection took two stages. The first was the introduction of fluorescent primers (Smith, Hood, and colleagues at Caltech, 1986): four separate reactions were run, each with a primer labelled with a different fluorescent dye, and all four were loaded together onto a single gel lane. The instrument detected the colour at each band position as the fragments passed a laser. The four-colour four-tube approach was adopted by Applied Biosystems in the ABI 370A automated sequencer (1987), the first commercial automated DNA sequencer.

The second and more significant step was the development of dye-terminator chemistry (Prober, Trainor, Dam, Hobbs, Robertson, Zagursky, Cocuzza, Jensen, and Baumeister, Science 1987; commercially developed by ABI in the Big Dye Terminator series): each ddNTP is covalently labelled with a unique fluorescent dye, typically a rhodamine or dichlororhodamine derivative. The four terminators emit at distinguishable wavelengths (ddATP: blue, ~520 nm; ddCTP: black/dark blue, ~560 nm; ddGTP: yellow/green, ~580 nm; ddTTP: red, ~620 nm in the Big Dye Terminator v3.1 chemistry used on ABI 3130/3500/3730 platforms). All four ddNTPs and the unlabelled dNTPs go into a single reaction tube with template, primer, and a thermostable polymerase. The reaction generates a mixture of labelled termination products of all lengths and all base types from a single tube.

The mixture is then purified to remove unincorporated dye-labelled terminators (which contribute a high background fluorescence) using ethanol precipitation or Sephadex column clean-up, resuspended in formamide, heat-denatured, and injected into a capillary for electrophoretic separation.

Big Dye Terminator v3.1 (Applied Biosystems, Thermo Fisher Scientific) is the chemistry used in most forensic capillary sequencing workflows globally. The FBI Laboratory Procedures Manual, the ENFSI Best Practice Manual for Mitochondrial DNA Profiling, and the Scientific Working Group for DNA Analysis Methods (SWGDAM) Interpretation Guidelines for Mitochondrial DNA Analysis all specify cycle-sequencing conditions compatible with this chemistry. NATA-accredited (Australia) and NABL-accredited (India) forensic laboratories that perform mtDNA typing run Big Dye Terminator v3.1 or its equivalent on ABI 3130xl, ABI 3500xl, or ABI 3730xl capillary instruments.

In gel-based Sanger sequencing, termination products were separated on a 6% denaturing polyacrylamide slab gel at 50-55°C. The gel format was replaced in the mid-1990s by capillary electrophoresis (CE), first demonstrated in automated instruments by Huang, Sanders, and colleagues, and commercialised by Applied Biosystems in the ABI 310 (single-capillary, 1995) and ABI 3100 (16-capillary array, 1997), leading to the current ABI 3130xl (16-capillary), ABI 3500xl (24-capillary), and ABI 3730xl (96-capillary) platforms.

In CE sequencing, fragments are separated inside thin fused-silica capillaries (50 µm internal diameter, 36 or 50 cm effective separation length) filled with a replaceable linear polyacrylamide or POP (Performance Optimised Polymer) gel matrix. The sample is electrokinetically injected at the inlet end; a high-voltage field (12-15 kV) drives fragments through the polymer by size: smaller fragments travel faster. Near the detector window, a focused laser (488 nm argon-ion or solid-state laser) excites the fluorescent dye on each fragment as it passes. Four photomultiplier tubes (or CCD arrays in newer instruments) simultaneously detect emission in four spectral bands corresponding to the four dye channels.

The raw output is four overlapping temporal fluorescence traces. A spectral deconvolution step (the "matrix" or "colour separation" step) applies an instrument-specific spectral calibration matrix derived from pure dye controls to separate the four channels. The resulting individual-channel traces are then processed by the base-calling algorithm. Phred, the landmark base-calling software developed by Phil Green and Brent Ewing at the University of Washington in 1998, assigns each peak a quality score (Phred score) based on peak spacing regularity, signal-to-noise ratio, and inter-peak spacing. A Phred score of 20 means a 1-in-100 probability of an incorrect base call; Phred 30 means 1-in-1000. Most forensic and genomic sequencing pipelines require Phred ≥ 20 (ideally ≥ 30) across the entire reported read window.

A standard ABI 3730xl read on optimised Big Dye v3.1 conditions delivers 700-1000 high-quality bases per read. For forensic mitochondrial DNA typing, the target regions are the hypervariable regions HV1 (positions 16024-16365, a 341 bp window) and HV2 (positions 73-340, a 268 bp window) within the control region of the mitochondrial genome. Both regions fit comfortably within a single Sanger read from a 16-mer primer, and in most forensic labs each region is sequenced in both directions (forward and reverse) so that two independent reads can be assembled and compared at every position.

The Human Genome Project (HGP), initiated in 1990 and jointly coordinated by the US Department of Energy, the US National Institutes of Health, and a consortium of international partners including the Wellcome Sanger Institute (Hinxton, UK), BGI (Beijing, China), Génoscope (Paris, France), and smaller centres in the US, Germany, and Japan, published the first draft human genome sequence in February 2001 simultaneously in Nature (International Human Genome Sequencing Consortium) and Science (Celera Genomics, Craig Venter's competing private effort). The finished "gold standard" sequence, representing approximately 92% of the euchromatic genome, was published in 2004.

The HGP used Sanger sequencing as its primary chemistry, but at a scale and degree of automation that had not previously existed. Ninety-six-capillary ABI 3730xl instruments, each running approximately 1,000 samples per day, operated continuously at the Wellcome Sanger Institute's sequencing centre in Hinxton. The sequencing strategy used BAC-based clone libraries: genomic DNA was sheared or partially digested with BamHI, size-selected fragments were ligated into BAC vectors (the cloning chemistry is described in restriction enzymes, vectors and molecular cloning), individual clones were fingerprint-mapped to produce a minimal tiling path across each chromosome, and each BAC clone was then shotgun-sequenced to 8-10-fold coverage. Individual shotgun reads (average 700-750 bases per read at Phred 20) were assembled into contigs and scaffolds, then scaffolded onto the BAC physical map using clone-end sequences.

The forensic relevance is direct. Every STR locus on the CODIS 20 panel was mapped to a chromosomal position using the HGP reference sequence. Every primer pair in every commercial STR kit (GlobalFiler, PowerPlex Fusion 6C, Investigator 24plex QS) was designed by aligning to the GRCh38 reference genome (the successor to the HGP's assembly), checking for off-target sites, and verifying repeat structure. When a forensic examiner reports that suspect A matches crime-stain B at locus D3S1358 with alleles 15, 16, that chromosomal address exists because of the HGP, and the HGP was built by Sanger sequencing.

Similarly, the rCRS (revised Cambridge Reference Sequence) for mitochondrial DNA, the global reference against which all forensic mtDNA profiles are reported, was itself sequenced by Sanger method (Andrews et al., 1999) and is deposited in GenBank (accession NC_012920.1). Every base-position call in an EMPOP database entry, every forensic mtDNA report that states "the crime-stain profile differs from the suspect at position 16189 (T>C)", refers to a coordinate in a Sanger-derived reference. Database querying strategies for those coordinates are detailed in sequence alignment, BLAST, GenBank, MITOMAP and EMPOP.

Illumina sequencing-by-synthesis (SBS), introduced commercially with the Genome Analyzer in 2006 and maturing with the MiSeq (2011), NextSeq, HiSeq, and NovaSeq platforms, uses a fundamentally different detection paradigm. A DNA library (adapter-ligated fragments) is attached to the surface of a glass flow cell and amplified locally into clusters. Sequencing occurs by successive addition of four reversibly fluorescent 3'-blocked nucleotides; each cycle a laser image captures which colour cluster incorporated which base; the 3'-block is removed; the next cycle proceeds. Read lengths are shorter than Sanger (150-300 bp per read in paired-end mode on MiSeq) but the throughput is vastly higher: an Illumina MiSeq run generates 15 million reads per run (approximately 22-25 million bases per run total on MiSeq v2 reagents), versus 96 reads per 3730xl run.

Oxford Nanopore sequencing (MinION, released to early-access users in 2014, widely available from 2015) works by threading a single-stranded DNA (or RNA) molecule through a protein nanopore embedded in a synthetic lipid bilayer. Changes in ionic current as successive bases thread through the pore are decoded into a sequence using a recurrent neural network base-caller (Guppy, then Dorado). Read lengths are limited only by fragment length; reads of 1-4 Mb have been reported from native mammalian DNA. The MinION device weighs 90 g, connects via USB, and has been operated in the field at crime scenes and disaster sites.

In a modern forensic lab, all three technologies coexist with complementary roles.

Sanger remains the gold standard for targeted sequencing of defined regions (mtDNA HV1/HV2, confirmation of mixed-source signals, verification of NGS variant calls, sequencing of specific STR alleles at forensic confirmation loci). Its per-base accuracy (Phred 30-40 in the core read) exceeds Illumina SBS for single-base resolution. The ABI 3500xl is the instrument of record in FBI-accredited US labs, ENFSI-member EU state labs, and NABL-accredited Indian FSLs for mtDNA typing.

Illumina MiSeq / Verogen MiSeq FGx is used for massively parallel sequencing (MPS) of STR loci with simultaneous SNP typing (the FORENSeq DNA Signature Prep assay) and for whole-mitochondrial-genome sequencing. The MiSeq FGx and its Universal Analysis Software (UAS) bioinformatics pipeline have received validation at multiple UK, US, Australian, and European forensic labs. King's College London's Department of Analytical, Environmental and Forensic Sciences published a validation study in 2019; the Netherlands Forensic Institute (NFI) published comparable data from their MiSeq FGx in 2018.

Oxford Nanopore MinION is still largely in research and validation phases for forensic use but is attracting serious attention for field-deployable rapid sequencing, DVI triage, and long-read haplotyping of complex genomic regions. A 2022 study from Radboud University Medical Center in the Netherlands demonstrated MinION sequencing of forensic STR loci from buccal swabs within 4 hours of collection.

Mitochondrial DNA sequencing is the primary active forensic application of the Sanger method. When skeletal remains are too degraded for nuclear STR typing (the standard situation with remains more than a decade old, or with hair shafts lacking a root), the high mitochondrial copy number (200-2,000 copies per cell) allows recovery of at least partial sequence from bone powder and teeth. The laboratory workflow is: demineralise the bone with EDTA or HCl, extract the organic fraction with phenol-chloroform or silica column, quantify mitochondrial DNA with a qPCR targeting the HV1 region (the Quantifiler Trio kit's human mitochondrial component), amplify HV1 and HV2 with locus-specific primers, and sequence by Sanger dye-terminator chemistry.

The first forensic mitochondrial identification of note was the Romanov identification (1991-1998): Gill, Ivanov, Kimpton, and colleagues (Nature Genetics, 1994) used Sanger sequencing of mtDNA from exhumed remains to confirm the identities of Tsar Nicholas II, Tsarina Alexandra, and three of their children, corroborated by a reference sequence from Prince Philip, Duke of Edinburgh, a maternal-line relative. The results were presented to Russian federal investigators and formed part of the official identification.

In the US, the FBI Laboratory's mtDNA Unit performs Sanger-based HV1/HV2 typing on unidentified remains submitted through the National Missing and Unidentified Persons System (NamUs). In the UK, Forensic Science International (and previously the Forensic Science Service before its closure in 2012) used the same chemistry. In Germany, the BKA's DNA laboratory at Wiesbaden uses ABI 3130xl capillary sequencing for mtDNA casework. In India, the CFSL Hyderabad's DNA section has implemented mtDNA sequencing as part of its skeletal remains identification workflow for disaster victim identification support.

Sequencing verification of STR alleles is a secondary use: when an off-ladder allele (an allele not matching any rung of the commercial ladder) is detected in a capillary electrophoresis STR run, Sanger sequencing of the amplicon confirms whether the deviation represents a true variant repeat structure or a sequencing/amplification artefact. This verification step is mandated by SWGDAM guidance and by the ENFSI DNA Working Group's guidelines for reporting variant alleles.

Mitochondrial haplogroup determination (placing a mtDNA profile within the global phylogenetic tree maintained by Mitomap and Phylotree) also relies on Sanger sequencing of coding-region SNPs in addition to the control region. Haplogroup data adds statistical weight when control-region sequences are common (matching dozens of EMPOP database entries) and informs population-frequency interpretation in admissibility hearings.