How to Choose a DNA Sequencing Service: Sanger, NGS, Turnaround

A DNA sequencing service is a core facility or commercial laboratory that generates nucleotide reads from submitted samples so a lab can verify constructs, call variants, or assemble genomes. How to choose a DNA sequencing service depends on method, sample class, turnaround, read quality, data formats, and confidentiality terms.

A method mismatch or a data package that cannot be aligned to the design file produces delays and unusable traces. Match the service to the biological question, then lock sample prep and the return format.

Start From the Question, Then Choose Sanger or NGS

Sanger sequencing still fits single-target confirmation: one plasmid region, one PCR amplicon, one clone to accept or reject. A typical Sanger read covers up to roughly a kilobase from a primer, with a chromatogram a person can inspect base by base. It is a poor fit when you need many samples, whole plasmids with difficult repeats, or genome-scale variant lists.

Short-read NGS fits multiplexed libraries, whole-plasmid sequencing, genomes, exomes, and targeted panels. Long-read NGS fits assemblies and structural context that short reads break. Whole-plasmid NGS can replace primer walking on a multi-kilobase construct, but repeats, homopolymers, and mixed-clone populations can look clean in a consensus and wrong in the raw reads. Ask whether the product is a consensus FASTA, a pileup, or both.

Core Facility Versus Commercial CRO

A university or institute core is often close, familiar with local primers, and run on a recharge model. Queues and limited hours can offset that convenience. A commercial CRO is built for inbound kits, documented turnaround windows, and a broader instrument list. Neither type is automatically more accurate. Accuracy follows sample quality, chemistry, and the analysis the provider actually runs.

• Queue and hours: A core that batches Sanger twice a week can be slower than a CRO that runs daily, even if the chemistry is the same.
• Method menu: Confirm that plasmid NGS, amplicon NGS, Sanger, or long reads are in production, not listed as a future offering.
• Rerun policy: Ask what happens to failed traces or low-quality libraries, and whether a rerun needs a new sample.
• Scientific contact: A named sequencing or bioinformatics contact matters when a map and a FASTQ disagree.
• Data custody: Compare how long traces are kept, who can download them, and whether accounts are personal or lab-owned.

For IP-sensitive work, a commercial contract can be easier to amend than campus core policy. For low-volume academic cloning, the core drop box is often the entire workflow. Choose the operating model, not a reputation slogan.

Sample Class: Plasmid, Amplicon, and Genome

Providers do not treat every tube as the same reaction. Plasmid DNA, PCR amplicons, genomic DNA, and RNA-derived libraries have different purity, concentration, and adapter needs. Sending a plasmid miniprep into a genome pipeline, or genomic DNA into a Sanger primer reaction, is a common source of failed runs that the provider will classify as a sample-prep problem.

State the sample class on the order form in the same words the provider uses. If the insert is a CRISPR edit, the service still needs to know whether you are confirming a small amplicon or surveying off-targets; guide RNA and sequencing primer design belongs in the experiment plan before tubes leave the bench.

Turnaround, Read Quality, and What Failure Looks Like

Turnaround is not one number. It includes accessioning, optional library prep, instrument queue, analysis, and the rerun clock. A Sanger trace that arrives quickly with a dye blob over the junction is not a successful service. Ask for the quoted window for your sample class, whether library prep sits inside that window, and what happens when quality gates fail.

Read quality should be defined in files, not in adjectives. For Sanger, that means chromatograms you can open and a clear view of where the read becomes unreliable. For NGS, that means Q scores, coverage along the target, and whether mixed bases were hidden in a consensus. A mixed plasmid can look clonal in a polished FASTA; ask for raw reads or a chromatogram if clonality is the question. Do not accept a PDF screenshot as the only archive.

Sample Prep Requirements That Change the Result

Most failed Sanger runs start with residual salts, RNA, too little template, or a primer that does not bind the submitted molecule. Most failed NGS runs start with degraded DNA, leftover primers in an amplicon, or a concentration measured on a dye that also sees RNA. Treat the provider's concentration and purity window as part of the method.

• Plasmid preps: Reduce RNA and host genomic DNA if the chemistry is sensitive to those contaminants, and submit the requested concentration.
• Amplicons: Remove leftover primers and dNTPs; they become extra sequencing primers or library junk.
• Primers for Sanger: Match Tm and uniqueness to the actual molecule, including adapters. A pUC19 vector primer may not exist on a redesigned backbone.
• Genomic material: Record integrity, mass, and expected genome size or panel on the intake form so the library kit is not a surprise.

If the provider offers optional cleanup or primer design, decide who owns that step before the tube is shipped.

Data Return Formats That Reconnect to the Design File

A sequencing service is only finished when the lab can compare the reads to the intended sequence. For Sanger, that comparison is an alignment of the base calls, plus a visual check of the AB1, against the plasmid or amplicon map. For NGS, it is FASTQ plus whatever alignment or assembly was in the quote. Ask for open formats: AB1 or SCF, FASTA or FASTQ, BAM or CRAM if alignments were included, VCF if variants were included, and a methods note. A colored PDF is not an archive.

Sequence visualization and alignment in ZettaGene are the point at which a deliverable becomes a cloning decision. Store the raw files next to that decision in ZettaNote so the chromatogram is not stranded in personal email. A connected R&D workspace helps only because design, alignment, and the lab note need the same construct name.

IP, Confidentiality, and Sample Retention

Construct sequences, patient-derived material, and unpublished genomes are not generic reagents. Before you ship, read who may use the sequence, where compute happens, how long DNA and files are kept, and how accounts close when a student leaves. Academic cores often operate under university policy; CROs operate under a service agreement. Neither is confidential unless the document says so.

Ask for written terms on secondary use, subprocessors, geographic processing, and destruction of leftover DNA. If the sample is human, add consent and identifiability questions your institution requires. If the sample is a proprietary plasmid, a license for the provider to keep sequences is a different deal from deletion after delivery. Inside the lab, use lab-owned accounts and named constructs; personal folders and unsigned PDFs are how traces leave with a departing user.

FAQ

When should I use Sanger sequencing instead of NGS?

Use Sanger when you need to confirm a single known region, accept or reject a clone, or inspect a chromatogram at a specific junction. It is the usual choice for one plasmid insert, one PCR product, or a small number of colonies, provided a primer exists on the molecule you actually submitted. Use NGS when you need many samples, whole-plasmid coverage, a genome or transcriptome, or a variant list that Sanger cannot produce at that scale. Whole-plasmid NGS is a reasonable Sanger alternative when primer walking would take many reactions or when repeats sit between primers. The deciding input is the question and the sample class, not a preference for a platform brand.

Should I use a university core facility or a commercial sequencing CRO?

Use the core when proximity, a drop box, and a recharge account match your volume, and when the core actually runs the method you need on a schedule you can live with. Use a commercial CRO when you need a documented turnaround window, a broader menu such as plasmid NGS or long reads, or a contract you can amend for confidentiality. Accuracy is not owned by either category; it follows sample prep, chemistry, and the files you receive. Compare queue behavior, rerun policy, scientific contact, and data custody. A nearby core that batches once a week can be slower than a remote CRO that runs daily. Choose the operating model, then validate it on a real construct, not on a brochure.

What sample type should I send for plasmid versus amplicon sequencing?

Send plasmid DNA for plasmid confirmation and an amplicon for a PCR product. Do not send a colony or a crude lysate unless the provider explicitly accepts that input. Plasmid preps should meet the stated concentration and purity window and should not be dominated by RNA if the chemistry is DNA-specific. Amplicons should be cleaned of leftover primers, which otherwise become extra sequencing primers or library contaminants. For genomes, send intact genomic DNA at the mass the library kit requires, with the expected genome size or panel written on the form. Mislabeling the sample class is one of the most common reasons a run fails and then gets classified as a user-prep problem.

What file formats should a DNA sequencing service return?

For Sanger, require chromatograms (AB1 or SCF) plus text base calls, not only a PDF picture. For NGS, require FASTQ as the minimum raw deliverable, and BAM/CRAM or VCF only if alignment or variant calling was part of the quote. Add a methods note that records chemistry, instrument class, and analysis version so you can interpret the files later. Open formats matter because you need to realign the reads to the design when a clone is disputed. A colored report that cannot be reopened in ordinary tools is not an archive. Keep those files with the construct name used in the lab, not under a provider job number that nobody will remember.

How do I judge sequencing quality without comparing list prices?

Judge the files and the failure policy. For Sanger, open the chromatogram and look at peak shape, the length of reliable sequence, and whether mixed peaks appear at the first position that should be clonal. For NGS, look at Q scores, coverage along the target, and whether a consensus hides a second species. Ask what the provider does with failed traces, low-diversity libraries, or contaminated plasmids, and whether a rerun needs new DNA. A cheap reaction that returns no usable chromatogram is not inexpensive. A more complete package that includes raw files and a defined rerun path is easier to defend in a cloning record, regardless of how the quote is structured.

What confidentiality terms should I require from a sequencing provider?

Require a written statement that submitted DNA and derived sequences will not be used for the provider's own assay development or for any secondary purpose without permission. Ask how long leftover DNA and files are kept, who can download them, where compute runs, and how access is revoked when a user leaves the lab. For human samples, add identifiability and consent requirements from your institution. For unpublished constructs, treat sequence as IP and reject vague language that grants the provider a license to retain data. Academic cores and commercial CROs both need this review; campus policy is not the same as a service agreement, and neither is confidential unless the document says so in operational detail.

Conclusion

How to choose a DNA sequencing service is a fit problem: Sanger versus NGS, core versus CRO, plasmid versus amplicon versus genome, then turnaround, readable quality metrics, open data formats, and terms that match the risk of the sample. The clone is not confirmed until the reads sit on the design file and the raw chromatogram or FASTQ is stored with that decision. To align sequencing files to plasmid maps and keep the comparison with the lab record, use ZettaGene in Zettalab's molecular biology tools.