From Swab to Sequence: How Viral Genomes Are Read

Every time a new variant is described, the starting point is the same: a sample taken from a person, a lab that extracts the virus's genetic material, and software that turns raw signals into a readable genome. This article walks through that pipeline in plain terms, so the phrase "the variant was identified by sequencing" means something concrete.

What Is Being Read

SARS-CoV-2 is an RNA virus. Its genome is a single strand of RNA roughly 30,000 letters long, one of the larger genomes among RNA viruses. Those letters, or nucleotides, encode the proteins the virus needs, including the spike protein that sits on its surface. Sequencing means determining the order of those letters as accurately as possible.

Most sequencing machines read DNA rather than RNA, so the first job is to convert the viral RNA into a DNA copy. This step uses an enzyme called reverse transcriptase and produces what is known as complementary DNA, or cDNA.

Step One: Collecting and Extracting

A sample usually begins as a nasal or throat swab collected for diagnostic testing. Samples that test positive, particularly those with a strong signal indicating plenty of viral material, are good candidates for sequencing. In the lab, the RNA is extracted from the sample and separated from proteins, salts and other material that could interfere with later steps.

The quality of this starting material matters. Samples with very little virus often produce incomplete genomes, with gaps where the sequence could not be determined.

Step Two: Amplification

Even a strongly positive sample contains far less viral genetic material than a sequencer needs. A widely used solution is amplicon sequencing. Scientists design many short primer pairs that tile across the whole genome, then use PCR to copy each overlapping segment many times. When the segments are later stitched together, they cover nearly the entire genome.

This approach is efficient and inexpensive, but it has a known weakness. If a new mutation falls exactly where a primer binds, that segment may copy poorly and leave a gap. Primer sets are periodically updated as the virus changes, which is one reason sequencing protocols are revised over time.

An alternative, metagenomic sequencing, reads all the genetic material in a sample without targeting the virus first. It is useful for discovering unexpected pathogens but generally needs more sequencing effort per genome.

Step Three: The Sequencing Run

Two broad families of technology are common in viral genomics:

• Short-read sequencing produces very large numbers of short, highly accurate reads. It is a workhorse for high-throughput public health labs.
• Long-read sequencing, including nanopore-based devices, reads longer stretches of genetic material and can be portable. Its per-read accuracy has historically been lower, though consensus methods compensate well for viral genomes.

Labs often add short barcode tags to each sample so that many samples can be pooled into a single run and separated again by software afterward.

Step Four: Building a Consensus Genome

The sequencer does not output a finished genome. It produces millions of fragments that must be aligned to a reference genome, a well-characterized earlier sequence used as a map. At each position, software looks at all the reads covering that spot and calls the most common letter. The result is a consensus genome representing the dominant virus population in that sample.

Positions with too few reads are marked as unknown, usually with the letter N. Quality-control steps flag genomes with too many unknown positions, unexpected frameshifts, or patterns suggesting contamination or a mixed infection.

Step Five: From Genome to Lineage

Once a consensus genome passes quality checks, it can be compared with large collections of other genomes. Software assigns it to a lineage and places it on a phylogenetic tree alongside its closest relatives. The naming and tracking side of this process is covered in our article on how scientists name and track new virus variants.

Genomes, along with basic metadata such as collection date and region, are often shared in public databases so researchers elsewhere can compare them. Personal identifying details of the patient are not part of this genomic record.

Limits Worth Knowing

Sequencing is powerful but not a complete census. Only a fraction of positive samples are ever sequenced, and that fraction varies between countries and over time. Regions with fewer sequencing resources can be underrepresented, and changes in testing behavior affect which infections are sampled at all. Delays between sample collection and data sharing also mean that genomic data describe the recent past rather than the present moment.

These limits do not make the data unreliable; they mean the data should be interpreted with sampling in mind.

Key takeaways

• Viral RNA is converted to DNA, amplified and then sequenced.
• Amplicon methods tile the genome with overlapping segments; new mutations can occasionally cause gaps.
• A consensus genome reflects the most common sequence at each position in a sample.
• Genomic data are a sample, not a full count, and should be read with that in mind.

Educational information only, not medical advice. For questions about testing, vaccination or treatment, follow the advice of a qualified healthcare professional and your local public health authority; in an emergency call your local emergency number.