What to check before a sequencing run

What to check before a sequencing run is a short list of evidence, not a feeling that the sample "looks fine". The instrument will read whatever molecules you give it. If the tube contains the wrong sample, the wrong length, too little sequenceable DNA, or two libraries wearing the same index, the files can still look busy. This page is a research checklist for Sanger sequencing and for short-read library runs. It is not a kit insert and not a clinical protocol. The wider path from library construction to reads is in next-generation sequencing from library to reads.

The decision the checks support

A run is justified when three statements are true. You know which biological question the reads are allowed to answer. You know the molecule in the tube is the one that chemistry can read. You know a failed control will be recognised as a failure. Coverage targets, variant claims and transcriptome counts all sit downstream of that gate.

Sanger sequencing reads one purified template with one primer, using chain-terminating synthesis and a capillary trace. A parallel short-read run reads millions of fragments that already carry adapters. The checks overlap, and they are not identical. A plasmid that is perfect for Sanger can be the wrong input for a whole-genome library, and a pooled library can be useless for a single-primer trace.

What amount, purity and size are actually saying

Absorbance at 260 nm is a convenient estimate of nucleic acid because the bases absorb there. It is also easy to misread. Free nucleotides, residual RNA, and some extraction chemicals contribute. The ratios of absorbance at 260 nm to 280 nm, and at 260 nm to 230 nm, are clues about protein and about some contaminants. They move with buffer and with pH. Treat a textbook ratio as a prompt to look harder, not as a pass stamp.

A dye that fluoresces when it binds double-stranded DNA, or single-stranded DNA, or RNA, answers a narrower question. When the dye reading and the absorbance reading disagree, believe that they disagree. The gap is often leftover nucleotides, a mixture of DNA and RNA, or a buffer that the spectrophotometer is scoring. Loading calculations that ignore the gap produce under-clustered or over-clustered runs.

Size matters because mass is not molecule count. A nanogram of a short fragment contains more ends than a nanogram of a long fragment. Clustering and emulsion-style chemistries care about ends. Record a size distribution from a fragment analyser, a capillary electrophoresis trace, or a gel that can actually resolve the range you claim. Adapter dimer is the small peak or band that competes efficiently and then wastes reads. If it dominates the trace, more sequencing does not repair the library.

For Sanger sequencing, size is the template itself. A PCR product needs one band that matches the design, or a cleanup story you can defend. A plasmid needs a prep that is not a smear of genomic DNA. The primer must match a site on that molecule. Mixed templates produce overlapping peaks, and no software preference will turn that into a clean consensus.

Library preparation details that the sheet must carry

Library preparation is a chain of irreversible choices: how the DNA or RNA was fragmented, which adapters were ligated, how many amplification cycles were used, and which indexes were assigned. The run sheet should name those choices in words a second person can follow. "Standard kit" is not a method.

Indexes are sample identity after the tubes are pooled. Two samples with the same index, or with indexes the analysis will treat as the same, cannot be separated later. Low-diversity index sets can also confuse instruments that expect base balance in the barcode read. Check the combination against the chemistry you will actually use, on paper, before the pool is made.

Amplification cycles are a quality choice. Extra cycles can rescue a low-input library and can also duplicate fragments until the reads look deep and the unique molecules are few. If the protocol you followed has a cycle recommendation tied to input mass, record the input you really had, not the input you hoped for.

A reference sequence belongs in the plan as well. NCBI GenBank is one public place to identify an accession. Name the assembly or the plasmid map the analysis will use. A later argument about a variant is often an argument about which reference was assumed.

Where the path branches

If identity is uncertain, stop. Relabeling after pooling is not a bioinformatics step. If amount is below the chemistry's stated input, either concentrate by a method that does not destroy the library, or accept a different assay. If the size trace shows a dimer peak that is a large share of the material, clean the library or rebuild it. If a Sanger control template of known sequence fails on the same day, the instrument or the premix is the first suspect, and the precious samples wait.

A spiked control library of known sequence, used in the way that chemistry describes, can separate a machine problem from a sample problem. It does not rescue a pool with collided indexes. A no-template library control, taken through preparation, tells you whether adapters and aerosols created a library from nothing.

How read quality fits after the run

Read quality is the next conversation, and it should be planned before the run so nobody is surprised. Tools such as FastQC summarise per-base scores, adapter content, duplication and odd sequence composition. Use them. Then ask the biological question with the reference you named. A drop in quality at the end of a read can be chemistry. A whole lane of the wrong organism is a sample-swap problem that quality scores will not announce in a headline.

Published technology notes, including the Illumina overview of sequencing by synthesis, describe one instrument family. Other platforms read different signals and ask for different library checks. Copy the checklist that belongs to the instrument you will use.

Rooms, heat and interrupted power

Libraries and enzyme mixes follow the storage on their own labels. In a hot building, a bench that a protocol calls "room temperature" may be warmer than the city where that protocol was written. Seal plates so humidity does not change the volume while a tray waits for a delayed run. If a power cut stops a sequencer mid-run, treat the partial data as a failed or incomplete run until the operator and the analysis record say otherwise. Do not silently merge it with a later repeat.

When libraries travel, the handoff matters as much as the preparation. Dry ice or cold packs only help if someone records that they were still cold on arrival. A warm shipment is a reason to repeat the size and amount checks, not a reason to hope.

What to send with a sequencing enquiry

EVRINTH can take a sourcing question. The useful note names the organism, the molecule (genomic DNA, amplicon, plasmid, or library already made), the read length and pairing you need, the reference accession, any index scheme already used, and the decision the data must support. Point to the genomics and sequencing catalogue for reagent and consumable classes, and to the genomics research pathway when the run sits inside a larger study.

The whole-genome sequencing reference and the Sanger DNA sequencing reference are independent method pages. Use them to frame questions. Ask whether a quotation is possible through the quote request. Do not read either page as a statement that EVRINTH operates that assay or holds a particular kit. A family name on a catalogue line is not a recipe and not evidence about a lot.

Safety

Sequencing reagents include irritant dyes, solvents and, on some instruments, high-voltage or laser components that belong to trained operators. The biological risk sits in the sample, not in the FASTQ file. Unknown human material, cultured pathogens and environmental samples follow your institutional biosafety rules. This article does not assign a containment level and does not authorise diagnostic reporting.