In biology, race is roughly analogous to the idea of subspecies, and the argument over its utility for humans has some history. After championing the concept as useful for categorizing human variation in the mid-20th century, the evolutionary biologist Theodosius Dobzhansky began to see that scientists needed a more precise way to talk about human diversity, and said as much. The race concept had been used to justify so much cruelty and oppression that its value, he came to realize, was questionable.

In the decades since, others have reiterated those points. What’s new today is that modern genetic science has revealed just how arbitrary the old race categories — Negroid, Caucasoid, Mongoloid and so on — really are. Yes, there is variation in the human family, but there are few sharp divides where one set of traits ends and another begins. Rather, traits exist in gradients, reaching high frequency in some populations and lower frequency in others. As the geneticist Sarah Tishkoff of the University of Pennsylvania reminded me, human beings are too young as a species, too promiscuous and full of wanderlust, always moving and mixing, for the kind of separation and differentiation that would cause true speciation to have occurred.

The truth is that there is sometimes more variation within what we call races than there is between them. A study led by the geneticist J. Craig Venter highlights this point. He compared his genome to that of James Watson, one of the two scientists who discovered the double helix shape of DNA. Dr. Venter focused on six genes that affect how we respond to antidepressants and antipsychotics, among other drugs. Dr. Venter and Dr. Watson are “Caucasian.” Dr. Venter had two copies of a “Caucasian” variant of one of these genes, but Dr. Watson carried a version more common in East Asian populations. Judging by self-reported race, Dr. Venter noted, a doctor might prescribe Dr. Watson the wrong dose of medicine.

The takeaway from studies like this is that rather than relying on race, doctors should focus on the genes important to whatever puzzle they face — an approach often called “precision” or “personalized” medicine. The idea is that tailoring treatment to the patient’s genotype, not to skin color or hair texture, would improve outcomes.

Consider the case of kidney disease. Scientists have found that African-Americans fare worse than whites when it comes to this illness. The assumption had long been that some environmental factor explained the difference. But in recent years, scientists have linked certain variants of a gene called APOL1 to worse kidney-related outcomes. Those variants are enriched in people of African ancestry. Girish N. Nadkarni, a kidney specialist at Icahn School of Medicine at Mount Sinai in New York City, explained to me that scientists think this may be because those variants protect against the sleeping sickness endemic to some parts of Africa.

But not all African-Americans carry the variants. That is important because, in an attempt to account for the higher risk of kidney problems among African-Americans, kidney donor registries include information about race. And kidneys from “black” donors might be discarded on the assumption that they are more prone to failure. But by looking directly at whether patients and donors carried this gene variant — in particular, two copies of it — scientists at the University of Alabama at Birmingham recently found they could theoretically improve transplant outcomes. Kidneys donated from black donors that might otherwise have been discarded could, with genotyping, go to patients who needed them.

Not everyone agrees that it is possible or even desirable to completely scrap the race concept. Alex Lickerman, founder of ImagineMD, a medical concierge service in Chicago, cites the example of prostate cancer. For unclear reasons, African-Americans have a higher risk than whites. One test for the cancer, which looks at prostate-specific antigen, is controversial because it can yield false positives. Some recommend against using it at all.