Here’s what you’ll learn when you read this story:

• An algorithm scanned 78 million genetic variants and found patterns that shouldn't exist.
• Most human DNA tells a familiar story—but some of it defies explanation.
• Researchers discovered genetic fingerprints that don’t match any known chapter of human history.

Homo sapiens are one species. There are no subspecies of humans, and our last relatives in the genus Homo went extinct around 40,000 years ago. But what explains how similar we humans are to each other? And more importantly, what genetic variants make different human populations and individuals so unique?

Some of the things we inherit in our genes are completely random. This is known as genetic drift, a phenomenon that happens because some individuals leave behind more offspring than others do. Their genes are more likely to be spread throughout the population, but it’s not because natural selection decided that those genes necessarily offer any added benefits. In fact, it’s even possible that a few of those genes are detrimental. Genetic drift is part of evolution and something that defines the genetic fingerprint of a specific group, but it doesn’t work by promoting adaptations that will help our species in the future.

Genetic drift isn’t exclusive to humans, but a team of researchers from the Institute of Statistical Science in Taipei, Taiwan, created an algorithm to figure out how it affects the frequency of alleles in human populations. Alleles are alternative versions of a given gene that come about through mutations. The team accessed data from the 1,000 Genomes Project, a database of single nucleotide polymorphisms (SNPs), which are common genetic variations caused by a difference in a single nucleotide within a DNA sequence. The researchers then used their algorithm to analyze recurring patterns of alleles in populations from different continents, as well as in groups within those populations.

“At the global scale, genetic differences between subjects faithfully reflect the well-known history of human population migration and mixing,” they said in a study recently published in Scientific Reports. “In contrast, when examining the allele frequency patterns of individual loci, we discovered many minor yet non-negligible…evolutionary trajectories in the human genome.”

Out of 78 million SNPs, the researchers found 71 patterns that tell the ancient history of how allele frequencies ended up being arranged in particular ways among various groups. It was hardly surprising that over 90% of such variants occurred with more or less the same frequency no matter what continent the population was from. The locations of most allele frequency patterns are scattered throughout chromosomes and are usually found together in “hot spots.” These narrow segments of DNA are linked to gene functions and observable phenotypical (body) traits.

Then the team used what they called a local ancestry inference algorithm to look deep into chromosomes and determine what ancestral groups an individual is descended from. The results from this algorithm were then checked against data from 1,000 Genomes and Human Genome Population Diversity (HGPD) data. It turned out that most allele patterns observed were synonymous with simulations that highlight the randomness of genetic drift. But there were also minor differences that suggest drift isn’t the only evolutionary process influencing the genetic profile of an individual or population.

Many findings were consistent with the migration of human ancestors out of Africa and into Europe, East Asia, and South Asia. Allele frequencies in African populations were found to be distinct from those of Eurasians in 1.9 million locations on their DNA. The separation of African and East Asian populations from populations in Europe and South Asia was evidenced by variants in 570,000 locations. This reflects the known history of human migration, which has Eurasians as the first to break from their African ancestors and trek through Europe and East Asia. South Asian populations arose from admixture between groups from Central Asia and migrants who first traveled from West to East Asia before heading south.

But the algorithm also turned up patterns that defy the accepted migration narrative. In 321,219 locations on DNA, allele frequencies of African and East Asian populations were more similar to each other than to those of Europeans and South Asians—the opposite of what the known migration timeline would predict. The researchers found that genetic drift likely explains many of these anomalies: random fluctuations in the allele frequencies of a small founding population can cascade through its descendants, producing arrangements that look historically “wrong” but are simply the residue of chance.

In at least one dramatic case, however, natural selection appears to be the culprit. One allele pattern, in which population frequencies are scrambled into a mosaic that ignores continental boundaries entirely, appeared nearly 70 times more often in real human data than drift alone would produce. It was concentrated almost exclusively on a family of immune-system genes on chromosome 6 called Human Leukocyte Antigen, or HLA. These genes build the proteins your body uses to identify invaders like viruses and bacteria. To do that job well, they need to stay wildly diverse. The more versions that exist, the more threats a population can recognize. So natural selection actively keeps them varied, no matter where on the planet a population lives. In other words, the immune system’s need to fight an unpredictable world has, at these specific locations, overwritten the genetic record of migration history.

“It is unclear whether the correlations between the evolutionary patterns and gene functions [or] phenotypes are mere coincidence or have causal implications,” the researchers said. “Nevertheless, our work sketches the landscape of such correlations on the human genome.”