Amplicon sequencing for a defined region

Amplicon sequencing reads the molecules your primers were willing to copy, at a depth that can make any of those molecules look important. The decision it supports is narrow and powerful: what alleles exist inside this defined region, in this mixture, above the error floor of this PCR. It is a bad genome. It does not become a genome because the read count is large. The instrument path is still next-generation sequencing from library to reads. One product read as peaks, without a flow cell, remains Sanger sequencing for a single amplicon.

Use an amplicon when the question is already a locus. A gene panel with a known coordinate list, a marker such as a ribosomal region, or the cut site of a genome edit are honest uses. A fishing trip for anything unexpected in the organism is not. Unexpected sequence outside the primers is invisible, and unexpected sequence inside them is still a PCR product with PCR's habits.

Primers are the universe

Polymerase chain reaction copies the span between two primers. The physical picture is the same cycle of denaturation, annealing, and extension described in how polymerase chain reaction works, repeated until the product dominates the tube. Sequencing that product tells you about the product. If a rearrangement deletes a primer site, that allele yields no reads and the other allele fills the file. If a SNP sits under the 3-prime end of a primer, polymerase may refuse that allele. The site then looks homozygous. Allele dropout is a false genotype produced by a successful neighbouring reaction.

Design primers on a named build, and check where else they might anneal. NCBI Primer-BLAST is the public check. A primer that also matches a paralogue will sequence both and the variant caller will report a heterozygous-looking mixture that is really two loci. A primer pair copied from a paper still needs that check against the build you will report, because reference genomes and why the build matters applies to primer coordinates as much as to variants.

Panel designs multiply the issue. Each tile has its own efficiency. GC-rich exons amplify late or not at all. A tiny exon may not fit a primer pair without walking into the next gene. The coverage plot across a panel is a picture of primer success. Reading it as a picture of copy number is how amplicon data invent deletions. There is a separate argument for capture-based exomes; this page is the PCR version of a defined region, and the comparison with genomes is in whole genome versus whole exome concepts.

From a PCR tube to a barcoded library

Two practical shapes exist. In a two-step PCR, the first primers amplify the target and carry tails, and a second reaction adds the full adapters and the sample indexes. In a one-step layout, the target primers already wear their indexes, which is simpler and costs a primer for every sample and locus combination. Either way the index rules in adapters indexes and barcode hopping apply. Unique dual indexes matter more, not less, on amplicons, because a hopped amplicon is a perfect-looking insert of the wrong sample and the read quality will not confess.

Cleanup between steps removes leftover primers that would otherwise prime the next round or form dimers. A dimer that survives will cluster and can dominate. Size-check the product. A single band at the designed length is the library you meant. A smear is a mixture of sizes the sequencer will not sort into "the" amplicon for you. Gel and cleanup thinking from agarose gel electrophoresis for DNA still applies before you spend a flow cell.

Unique molecular identifiers, when the primer design includes them, let you collapse PCR copies back toward starting molecules. They are a reagent design, not a software theme you add later. Without them, duplicate marking on amplicons is weak, because every genuine molecule of a short identical insert looks like a duplicate. Depth on an amplicon is therefore a poor proxy for independent observations. Say so when you report an allele fraction.

The polymerase's error rate is the floor under rare variants. An early error is amplified into a family of reads. High depth makes that family obvious and does not make it biological. A claimed allele at a fraction near the polymerase's error, seen in one PCR, is a hypothesis. A second independent PCR that sees the same allele is the start of evidence. The no-template control must be sequenced or at least shown blank, because a contaminant amplicon sequences at the same handsome depth as a real one. PCR controls and contamination control is the discipline those wells belong to.

Branch when the pattern is uneven or empty

If every sample, including the no-template control, shows the product, stop. You are sequencing a contaminant. More reads will characterise the contaminant with great precision.

If one amplicon in a panel is absent from every sample, including a positive control, the primers or the PCR condition failed. That is a design or a reagent fault. Do not impute the missing exon as a deletion in the cohort.

If one amplicon is absent from one sample and present in the others, you have either a true deletion of the primer sites, an allele dropout, an empty well, or a failed aliquot. Distinguish them with a second primer pair placed elsewhere in the same region, or with a method that does not use these primers. One missing tile is not a copy-number call.

If depth varies a hundred-fold across tiles that should be single-copy, believe the primers first. Normalise only when you understand what you are dividing by. A sample with globally low depth is a library or loading problem. A sample with one explosive tile is a jackpot amplification. Coverage depth is not the same as accuracy is the longer version of that warning, and it is sharper here because the PCR manufactured the depth.

Chimeras, jackpots, and a region pretending to be a genome

Chimeras are the failure mode of mixed templates. Incomplete strands anneal across templates. The resulting read is locally real and globally false. Marker-gene surveys suffer most, which is why 16S profiling and its taxonomic limits treats chimera screening as part of the method rather than as an optional filter. A gene panel from one diploid genome makes fewer cross-template chimeras and can still make them between paralogues.

Jackpot tiles steal the run. The sequencer's fixed read budget goes to the amplicon that amplified best, and the weak tile falls below the depth at which an allele fraction means anything. Balancing primer concentrations is empirical. It belongs to the kit or to a pilot, not to a hope that analysis will reweight a tile that has twenty reads.

The genome-shaped over-reach looks like this. A laboratory sequences twelve exons, finds no variant, and writes that the gene is intact. Introns, promoters, and the exons the primers missed were never observed. A structural variant that removes an exon and both primer sites looks identical to a failed PCR in that one sample. Phrase the negative as "no variant was called inside the amplicons that succeeded".

The upstream sample can be infectious

Amplicons are nucleic acid, and the culture, swab, or tissue before extraction may be infectious. PCR does not grant a biosafety category. The institutional decision covers how the sample is opened and whether the amplicon may move to a sequencing bench. This guide is not a diagnostic approval. A pathogen amplicon sequenced for research is not a clinical identification, even when the primer name matches a published assay. Keep the no-template and, where the question is a pathogen, a negative extraction control. A positive research control of a pathogen amplicon is also a contamination source for every future PCR in that room.

Primer stocks, heat, and a written design

Working primer stocks left on a warm bench, or freeze-thawed through a week of unstable cold storage, become a different reagent. Aliquot, and record the lot. In a humid week, an unsealed plate of a long PCR concentrates and the later cycles misbehave; seal and spin. Write the primer sequences, the build they were designed on, and the expected length into the specification that travels with the plate. A collaborator who receives only a gene name cannot tell dropout from a real absence. If a power cut reboots a cycler mid-programme, treat the plate as unfinished and rerun the controls before you sequence it. A half-cycled amplicon is a chimera factory.

What the enquiry needs

State the coordinates, the build, the primer sequences or a locked panel version, the polymerase class, whether indexes are unique dual indexes, and what a missing tile should mean. Say whether the question is a majority allele, which Sanger might answer more directly, or a mixture that needs counts. Include the no-template plan.

A defined-region design can be discussed as sequencing work against the whole-genome sequencing enquiry reference only when you are explicit that the deliverable is amplicons, not a genome. The single-product alternative can be discussed against the Sanger DNA sequencing enquiry reference. Primer and enzyme classes sit in the genomics and sequencing catalogue, and the study context is the genomics research pathway. Put the window and the controls in the quote request. What to check before a sequencing run includes the size and the blank. Commissioning a sequencing or proteomics study is how those stay in the statement of work.