Sanger Sequencing: Principles, Reading a Trace, and When to Use It vs. NGS

Sanger sequencing (chain-termination sequencing) is still the workhorse method for confirming that a specific, known stretch of DNA — a plasmid insert, a CRISPR edit site, a single PCR amplicon — reads the way you think it does. It is not obsolete; next-generation sequencing (NGS) did not replace it, it took over a different job. For targeted, single-sample, human-readable confirmation, Sanger remains faster to set up, cheaper per reaction at low sample numbers, and easier to interpret without a bioinformatics pipeline. This guide covers the chemistry, how to read a chromatogram (the actual diagnostic skill), a symptom-based troubleshooting table, when to choose Sanger over NGS, and the research-administration mechanics — core facility submission, cost structure, and sequence deposition as a publication requirement.

The chain-termination principle

Sanger sequencing, first described by Frederick Sanger and colleagues in 1977 (Sanger, Nicklen & Coulson, “DNA sequencing with chain-terminating inhibitors,” PNAS 74(12): 5463-5467), works by controlled, random termination of DNA synthesis.

A sequencing reaction contains a single primer, template DNA, DNA polymerase, the four normal deoxynucleotides (dATP, dCTP, dGTP, dTTP), and a small proportion of the four dideoxynucleotides (ddATP, ddCTP, ddGTP, ddTTP). A dideoxynucleotide lacks the 3′-hydroxyl group that DNA polymerase needs to attach the next base — so wherever a ddNTP is incorporated instead of the normal dNTP, synthesis stops permanently at that position. Because ddNTP incorporation happens stochastically across many millions of template copies in the same tube, the reaction produces a population of fragments of every possible length, each one terminating at a different base and each one tagged with a fluorescent dye specific to which of the four ddNTPs stopped it (modern automated, “dye terminator” sequencing labels the four terminators with four distinct fluorophores in a single reaction, rather than running four separate reactions as in the original manual method).

Those fragments are then separated by size using capillary electrophoresis (CE): an automated sequencer (commonly an Applied Biosystems/Thermo Fisher instrument such as the 3500 or SeqStudio series) injects the reaction into a narrow capillary filled with a sieving polymer and applies an electric field. Because DNA is uniformly negatively charged, all fragments migrate toward the anode at a rate governed almost entirely by length — shorter fragments move faster. A laser near the far end of the capillary excites the fluorescent tag on each fragment as it passes a detection window, and the resulting color/intensity signal over time is reconstructed, one base at a time, into the trace you actually look at: the chromatogram (also called an electropherogram) — a series of four-colored peaks, one color per base, with the base-calling software (in ABI’s ecosystem, historically KB Basecaller) assigning the most probable base at each position based on peak shape, spacing, and height.

Usable read length, and why the first bases are unreliable

A single, well-optimized Sanger reaction typically returns roughly 700-900 bases of high-quality, trustworthy sequence, with total raw read length extending somewhat further (often cited up to 800-1,000 bases) before quality drops off entirely. This is the single biggest practical constraint on the method: it is why Sanger is used for targeted confirmation of a known region rather than for assembling anything longer, and why sequencing a long insert requires either primer walking (a series of overlapping reads from staggered primer sites) or a switch to NGS.

The trace is not uniformly usable across its full length. Two things new users consistently get wrong:

• The first ~20-40 bases are typically unreliable and often unreadable. This is a “ramp-up” artifact of the chemistry and detection, not a sign anything went wrong: the earliest-terminating fragments are the shortest and migrate together in a tight, poorly-resolved cluster near the injection point, so peaks in this region are often uneven, overlapping, or missing. Design your primer, and place anything you actually need to confirm (a mutation site, a cloning junction), at least 30-50 bases downstream of the primer’s 3′ end, never right at it.
• Quality degrades again toward the far end of the read, past roughly base 700-900, as fragment resolution declines with capillary run time and signal-to-noise falls. This is the normal, expected tail-off of a healthy run, distinguishable from a genuine trace failure (see the troubleshooting table below) by the fact that it is gradual, not sudden.

Practical implication: if you need to confirm a feature near either end of your amplicon or insert, design a second primer that puts that feature in the high-quality middle of its own read, rather than relying on the edge of a single trace.

Reading the chromatogram: Phred quality scores and what a double peak means

This is the actual diagnostic skill the rest of this guide’s troubleshooting section builds on, and it is the thing most commonly under-taught: a Sanger result is not “pass/fail,” it is a trace you read.

Phred quality scores

Each base call is assigned a Phred quality score, an established convention (originally developed for shotgun sequencing base-calling and adopted throughout the field including ABI’s own basecallers) expressing the estimated probability that the call is wrong, on a log scale: Q10 = 1-in-10 chance of error, Q20 = 1-in-100, Q30 = 1-in-1,000, Q40 = 1-in-10,000. As a working rule, treat base calls at Q20 or above as trustworthy and be cautious about relying on anything below that threshold, particularly at a position you’re using to call a specific variant or mutation. Most trace-viewing software (Sequence Scanner, SnapGene, 4Peaks, Geneious, or the free NCBI/EMBOSS trace viewers) displays quality as a bar or shaded band under the chromatogram so you don’t have to read raw numeric scores.

Double and mixed peaks

A position where two colors/peaks appear stacked at similar height is the most common thing a new user doesn’t know how to interpret, and it has at least three genuinely different causes that call for different responses:

• A true heterozygous variant (in diploid template). If you are sequencing genomic DNA from a diploid organism directly (rather than a cloned, haploid plasmid insert), a clean, roughly 50:50 double peak at a single position, with clean single peaks on either side, is the expected signature of heterozygosity at that base — this is a real biological result, not an error.
• Mixed template. If double or messy multi-position peaks appear broadly across much of the trace (not confined to one clean position), the template itself is very likely a mixture — for example, a bacterial colony pick that wasn’t from a single clone, a plasmid prep contaminated with a second construct, or (in the CRISPR-editing context) a PCR product amplified from a mixed population of edited and unedited alleles, which is exactly the kind of mosaic signal you’d expect from an unsorted, non-clonal edited cell pool. The fix is to re-isolate a single clone or colony and re-sequence, not to trust the mixed read.
• A primer binding-site problem. If the trace looks clean and single-peaked for a stretch and then degrades into overlapping, out-of-register double peaks partway through — often with the two apparent “reads” seeming to be offset copies of each other — the primer is likely annealing at more than one site (a repeat region, a second homologous site in the template, or a primer that also binds vector backbone near an insert junction). This produces two overlapping reading frames from two different start points in the same capillary run, not a real biological mixture at the template level.

Distinguishing these three matters administratively as well as scientifically: a heterozygous call is a result to record, a mixed-template call means re-picking material before you draw any conclusion, and a primer-site problem means redesigning the primer rather than repeating the same reaction and expecting a different answer.

Sample and primer requirements

Exact concentration and volume requirements vary by core facility and by commercial provider (Azenta/GENEWIZ, Eurofins, and university cores each publish their own submission tables), so always check the specific provider’s current submission guidelines before preparing samples — the ranges below are typical, not universal, and using an out-of-range concentration is one of the most common reasons a run fails or comes back short. In general terms:

• Template type matters: plasmid/BAC DNA, purified PCR product, and bacterial colonies/cultures each have their own required concentration ranges and minimum volumes, because the facility is normalizing input DNA mass, not volume, into the sequencing reaction.
• PCR products should be cleaned up (column or enzymatic cleanup to remove unincorporated primers and dNTPs) before submission — residual primers from the PCR step are a common cause of poor or double traces, since they can act as unintended additional priming sites in the sequencing reaction itself.
• Primers are typically submitted separately, unlabeled, as standard desalted oligos at a facility-specified concentration; a single sequencing-grade primer is normally sufficient input for many reactions, so a small stock goes a long way. Facilities generally will not accept primers that also carry a 5′ modification unless the run is specifically set up for it.
• Template purity, not just quantity, drives results: residual salts, ethanol, phenol, or protein carryover from extraction depresses signal or introduces noise even when the DNA quantity itself looks correct on a Nanodrop/Qubit reading.

See our related guide on PCR reaction setup for amplicon preparation before submission.

Troubleshooting by symptom

Symptom	Likely cause(s)	What to try
No signal / blank trace	No template or primer added to the reaction; primer doesn’t match the template at all (wrong primer, wrong construct); failed PCR submitted without checking on a gel first	Confirm reaction setup; run a quick gel check on PCR product before submission; re-submit with confirmed template and correctly paired primer
Weak/low signal throughout	Template concentration too low; primer concentration too low; poor-quality or degraded template	Re-quantify template accurately (Qubit is more reliable than Nanodrop for this purpose) and resubmit within the facility’s target concentration range; check template integrity
Noisy baseline / high background	Template or primer contamination (salts, ethanol, protein carryover from extraction); non-specific primer binding; excess primer-dimer or unincorporated primer in an uncleaned PCR product	Re-purify template with a clean-up column; clean up PCR product before submission; check primer specificity against the template sequence
Sudden signal collapse partway through the read	A strong secondary structure (hairpin, GC-rich region) in the template stalling polymerase; a homopolymer run or repeat region; abrupt loss of resolution from a capillary/instrument issue on the provider’s end	Try a GC-rich-optimized chemistry or additive if the facility offers it; design a second primer to approach the problem region from the opposite direction; if isolated to one run, treat as an instrument artifact and re-submit
Mixed/double peaks throughout the entire trace	Non-clonal template (mixed colony pick, contaminated plasmid prep, unsorted mixed-allele CRISPR pool); two co-amplified PCR products of similar size	Re-isolate a single colony/clone; re-transform and re-miniprep; gel-purify the correct PCR band before re-submission
Short reads (drops out well before the expected ~700-900 bases)	Low template input; primer binding-site secondary structure; template degradation; a homopolymer or repetitive region collapsing the read early	Increase template concentration within the recommended range; redesign primer position; check template integrity by gel before resubmission

When Sanger is still the right tool — and when NGS wins

The decision is about what you already know and how many targets you have, more than accuracy — both methods are highly accurate when run correctly.

Choose Sanger when

• Plasmid or clone verification — confirming an insert sequence, orientation, and junction after cloning, where you already know the expected sequence and just need to confirm the construct matches it.
• CRISPR edit confirmation at a single, known locus in a small number of clonal lines — for a clean edited clone, Sanger directly reads the edit; for a mixed/unsorted pool, a double-peak pattern around the cut site is itself diagnostic of editing efficiency (and tools such as ICE or TIDE are built specifically to deconvolute that mixed Sanger trace into an indel-frequency estimate, without needing NGS).
• Confirming or validating a specific variant an NGS run already called — Sanger is the standard orthogonal confirmation method for a single candidate variant flagged by exome/panel sequencing, precisely because it’s an independent chemistry and platform from the NGS call it’s checking.
• A small number of targeted amplicons across a small number of samples, where per-sample turnaround and simplicity of analysis (a chromatogram you can read yourself, no bioinformatics pipeline) matter more than throughput.

Choose NGS when

• You need whole-genome or whole-exome coverage, or don’t yet know where in a large region a variant might be — Sanger only tells you about the specific stretch you chose to prime and sequence.
• You’re looking for unknown or novel variants across a large target region, rather than confirming a specific, already-suspected position.
• You need to detect low-frequency variants (e.g., subclonal tumor mutations, rare alleles in a mixed population) below roughly the 15-20% level — Sanger’s double-peak detection has a practical sensitivity floor well above what NGS depth-of-coverage can resolve.
• You have many targets or many samples at once — NGS’s per-base cost falls sharply with scale in a way a per-reaction Sanger workflow cannot match once you’re beyond a handful of targeted regions.

Core facility submission, cost, and turnaround

Most institutions route Sanger sequencing through an internal genomics/sequencing core facility rather than individual labs running their own capillary sequencers, and many of those cores now manage submission, sample tracking, and billing through platforms like the ones covered in our guide to core facility management software. A few things that matter for planning and budgeting a project around Sanger sequencing:

• Recharge/pricing structure. Cores typically bill per reaction (one primer, one template, one read), with a separate, usually lower, external/industry rate versus an internal institutional rate — check your own facility’s current published rate card rather than assuming a figure, since per-reaction pricing varies meaningfully by institution, read length tier, and whether you supply your own primers or purchase them through the facility.
• Turnaround. A same-day or next-business-day result for standard submissions is common at many university cores and commercial providers when samples are submitted before a facility’s daily cutoff time, though this varies by provider, batch size, and whether you’ve chosen a standard or premium/rush service tier.
• Grant budgeting. Because Sanger is billed per reaction rather than per project, a study involving repeated clone verification (a common step in molecular cloning-heavy grant-funded work) can accumulate a non-trivial number of individual line items over a project’s life — worth itemizing in a budget justification rather than folding into a single generic “sequencing” line, particularly for grants with detailed budget review.

Sequence deposition to GenBank as a publication requirement

Many journals require that any novel nucleotide sequence reported in a manuscript — including a newly verified plasmid construct, a novel gene variant, or a sequenced clone central to the paper’s findings — be deposited in a public sequence database (most commonly GenBank, NCBI’s primary nucleotide sequence archive, or one of its International Nucleotide Sequence Database Collaboration partners, EMBL-Bank/ENA and DDBJ) prior to or at the time of publication, with the resulting accession number cited in the manuscript. This is a standard element of a journal’s data-availability policy, distinct from — but often satisfied using the same underlying trace data as — the raw Sanger chromatogram file itself. For a research administrator or lab manager tracking a project through to publication, this means the Sanger confirmation run that validated a construct early in a project is frequently the same data that needs to be retrievable, and correctly attributed to the right author and repository record, at submission time. Building accession-number deposition into your project’s data-management planning at the point you first run the sequencing — rather than scrambling to reconstruct which trace corresponds to which construct months later during manuscript submission — avoids a genuinely common late-stage bottleneck.

Frequently asked questions

How accurate is Sanger sequencing?

Within its high-quality read window (roughly bases 30/40 through 700-900, at Phred Q20 or above), Sanger sequencing is generally regarded as the most accurate widely-used sequencing chemistry for a single targeted region, which is exactly why it remains the standard method for orthogonally confirming a variant an NGS run has already flagged.

Can Sanger sequencing detect a heterozygous mutation?

Yes — a clean double peak at a single position, of roughly equal height, against clean single peaks on either side, is the expected chromatogram signature of a true heterozygous variant. See the chromatogram-reading section above for how to distinguish this from a mixed-template or primer-binding artifact that can look superficially similar.

Why is the beginning of my Sanger trace unreadable?

The first roughly 20-40 bases after the primer are a normal “ramp-up” region of poor peak resolution inherent to the chemistry and capillary detection, not evidence of an error in your reaction. Design primers so anything you actually need to read sits well downstream of this zone.

Do I need NGS to confirm a CRISPR edit?

Not usually, for a single known locus in a small number of clonal lines — Sanger sequencing directly reads a clean edited clone, and for a mixed/unsorted pool, deconvolution tools such as ICE or TIDE can estimate editing efficiency directly from the mixed Sanger trace around the cut site. NGS becomes the better choice when you need to quantify low-frequency edited alleles precisely, screen many clones/samples at once, or check for off-target edits across multiple loci.

How long does Sanger sequencing take at a typical core facility?

Many university core facilities and commercial providers offer same-day or next-business-day turnaround for standard submissions received before a daily cutoff, though this varies by facility, current queue, and service tier — check your specific provider’s current published turnaround time.

What does a “mixed peak” across an entire trace usually mean, as opposed to one clean double peak?

A single clean double peak at one isolated position, on an otherwise clean trace, usually indicates a true heterozygous site. Double or overlapping peaks spread broadly across much of the trace usually indicate a non-clonal or contaminated template — most commonly a colony pick, plasmid prep, or PCR product that isn’t from a single, pure source — and calls for re-isolating material rather than trusting the read.