chromosomes.
Consider the following thought experiment. Say the hemochromatosis gene sits on chromosome seven, and the gene that governs hair texture—straight versus kinked or curly or wavy—is its immediate neighbor on the same chromosome. Now assume that somewhere in distant evolutionary history, the defective hemochromatosis gene arose in a man with curly hair. Every time this ancestral gene is passed from parent to child, the curly-haired gene travels with it: both are bound on the same chromosome, and since chromosomes rarely splinter, the two gene variants inevitably associate with each other. The association may not be obvious in a single generation, but over multiple generations, a statistical pattern begins to emerge: curly-haired children in this family tend to have hemochromatosis.
Kravitz and Skolnick had used this logic to their advantage. By studying Mormons in Utah with cascading, many-branched family trees, they had discovered that the hemochromatosis gene was genetically linked to an immune-response gene that exists in hundreds of variants. Prior work had mapped the immune-response gene to chromosome six—and so the hemochromatosis gene had to be located on that chromosome.
Careful readers might object that the example above was loaded: the gene for hemochromatosis happened to be conveniently linked to an easily identifiable, highly variant trait on the same chromosome. But surely such traits were fleetingly rare. That Skolnick’s gene of interest happened to be sitting, cheek by jowl, with a gene that encoded an immune-response protein that existed in many easily detectable variants was surely a lucky aberration. To achieve this kind of mapping for any other gene, wouldn’t the human genome have to be littered with strings of variable, easily identifiable markers—lamplit signposts planted conveniently along every mile of chromosome?
But Botstein knew that such signposts might exist. Over centuries of evolution, the human genome has diverged enough to create thousands of minute variations in DNA sequence. These variants are called polymorphisms—“many forms”—and they are exactly like alleles or variants, except they need not be in genes themselves; they might exist in the long stretches of DNA between genes, or in introns.
These variants can be imagined as molecular versions of eye or skin color, existing in thousands of varied forms in the human population. One family might carry an ACAAGTCC at a particular location on a chromosome, while another might have AGAAGTCC at that same location—a one-base-pair difference.I Unlike hair color or the immune response, these variants are invisible to the human eye. The variations need not enable a change in phenotype, or even alter a function of a gene. They cannot be distinguished using standard biological or physical traits—but they can be discerned using subtle molecular techniques. A DNA-cutting enzyme that recognizes ACAAG, but not AGAAG, for instance, might discriminate one sequence variant and not the other.
When Botstein and Davis had first discovered DNA polymorphisms in yeast and bacterial genomes in the 1970s, they had not known what to make of them. At the same time, they had also identified a few such polymorphisms scattered across human genomes—but the extent and location of such variations in humans was still unknown. The poet Louis MacNeice once wrote about feeling “the drunkenness of things being various.” The thought of tiny molecular variations peppered randomly through the genome—like freckles across a body—might have provoked a certain pleasure in a drunken human geneticist, but it was hard to imagine how this information might be useful. Perhaps the phenomenon was perfectly beautiful and perfectly useless—a map of freckles.
But as Botstein listened to Kravitz that morning in Utah, he was struck by a compelling idea: if such variant genetic signposts existed in the human genome, then by linking a genetic trait to one such variant, any gene could be mapped to an approximate chromosomal location. A map of genetic freckles was not useless at all; it could be deployed to chart the basic anatomy of genes. The polymorphisms would act like an internal GPS system for the genome; a gene’s location could be pinpointed by its association, or linkage, to one such variant. By lunchtime, Botstein was nearly frantic with excitement. Skolnick had spent more than a decade hunting down the immune-response marker to map the hemochromatosis gene. “We can give you markers . . . markers spread all over the genome,” he told Skolnick.
The real key to human gene mapping, Botstein had realized, was not finding the gene, but finding the humans. If a large-enough family bearing a genetic trait—any trait—could be found, and if that trait could be correlated with any of the variant