Sequencing explained: panels, exomes, genomes and variant classification

Sequencing DNA means determining the order of the four chemical bases that make up the DNA molecule.^[1] The original method, Sanger sequencingSanger sequencingAn earlier sequencing method, used by the Human Genome Project and still used for short pieces of DNA., helped scientists determine the human genetic code but is time-consuming and expensive; it is still used today for short pieces of DNA. Next-generation sequencingNext-generation sequencing (NGS)Newer, faster sequencing approaches that make it feasible to sequence large amounts of DNA, such as a whole exome or genome, in days to weeks. has made it feasible to sequence large amounts of DNA, taking days to weeks for a human genome.^[2]

Walter Gilbert and Frederick Sanger shared half of the 1980 Nobel Prize in Chemistry “for their contributions concerning the determination of base sequences in nucleic acids”.^[3]

Clinical laboratories perform genetic testing that spans single genes, gene panelsGene panelA test that looks at a selected set of genes, usually chosen for a particular condition or question., exomes and genomes.^[4] Whole exome sequencingWhole exome sequencing (WES)Sequencing only the exons. It can miss variants that lie outside them. reads the exonsExonThe protein-coding parts of a gene. Together they are thought to make up about 1% of the genome. — the protein-coding pieces of DNA, thought to make up about 1 percent of the genome — so it can miss variants outside them. Whole genome sequencingWhole genome sequencing (WGS)Sequencing that determines the order of all the nucleotides in a person’s DNA, both inside and outside genes. determines the order of all the nucleotides in a person’s DNA.^[2]

Looking at more DNA brings trade-offs. The significance of much of the variation found is not yet known, and testing can identify a variant linked to a different condition from the one being investigated — an incidental or secondary finding.^[2]

Sequencers produce many short readsReadA short stretch of DNA sequence produced by a sequencing machine. Reads are lined up against a reference sequence to find differences. that are compared with a reference genome. The number of reads overlapping a base is its depth, or CoverageCoverage (sequencing depth)How many reads overlap a base. Average depth is the number of reads × read length ÷ the size of the target; breadth is the share of the target covered at least a given number of times.. In theory, average depth is L × N ÷ G, where L is the read length, N the number of reads and G the length of the genome. Breadth of coverage is the percentage of target bases sequenced a given number of times.^[5]

Variant callingVariant callingFinding positions where the reads consistently differ from the reference sequence. is the process of identifying consistent differences between the reads and the reference. Its accuracy is affected by sequence quality, how evenly the target is covered and the statistical threshold used.^[5]

Germline variantsGermline variantA variant inherited from a parent through the egg or sperm. It is present in virtually every cell of the body. are inherited from a parent through the egg or sperm and are present in virtually every cell of the body. Somatic variantsSomatic variantA variant acquired during a person’s life in certain cells only, for example from copying errors or UV exposure. It is not passed to children. are acquired during a person’s life and occur only in certain cells; they can be caused by environmental factors such as ultraviolet radiation or by copying errors as cells divide, and they are not passed on to children.^[7]

Somatic variants in cancer cells are interpreted differently: the proportion of reads carrying a variant is highly variable and differences within a tumour can affect what a sample shows, so separate interpretation guidelines are needed.^[4]

In 2015 the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) published a joint standard for interpreting sequence variants in inherited, single-gene (Mendelian) conditions. It uses five terms: pathogenic, likely pathogenic, uncertain significance, likely benign and benign.^[4]

Evidence comes from sources such as population data, computational predictions, functional studies and how a variant tracks with a condition in families. Each pathogenic criterion is weighted as very strong, strong, moderate or supporting. As evidence evolves, earlier classifications may need to change.^[4]

The guideline recommends the neutral word “variant” over “mutation” and “polymorphism”, which can wrongly suggest harm or harmlessness. It is not intended for somatic variants, pharmacogenomic variants or complex conditions involving many genes.^[4]