Skip to main content

Introduction

Every time a cell divides, it copies all of its DNA, about six billion bases. The copying is remarkably accurate, and repair machinery fixes most of the mistakes that slip through, but not all of them. The outside world adds more: ultraviolet light and the chemicals in tobacco smoke can both damage DNA. Any difference in a DNA sequence is called a variant. You have already met two. F508del deletes three bases from CFTR, and G542X changes a single base, turning a codon into a stop.

Where a variant arises decides who inherits it. A variant present in an egg or sperm cell is passed on to the child and ends up in every cell of their body. These are germline variants, and they are the ones that run in families. A variant that arises later, in a single cell of the body, is passed only to that cell's descendants and never to the person's children. These are somatic variants. Most cancers are driven by somatic variants, like the second hit in retinoblastoma we discussed in Oncogenes and Tumor Suppressors.

Variants are everywhere. A typical person's genome differs from the standard human reference (which we define on the next page) at roughly four to five million positions. Most of these differences have little or no effect. A few are harmful, and fewer still are helpful. Over many generations, natural selection tends to keep the helpful ones and weed out the harmful ones, which is how species evolve.

For anyone working with genomic data, finding variants is one of the central jobs. Comparing a sequenced sample against a reference and listing the differences is called variant calling, and much of the rest of this guide builds toward it. This chapter covers the kinds of variants you will be looking for: