Skip to main content
Premium Trial:

Request an Annual Quote

African Population GWAS Reveals Ancestral Alleles Linked to Light, Dark Pigmentation

NEW YORK (GenomeWeb) – A new genetic study focused on populations in Africa has unearthed previously unappreciated skin pigmentation contributors — genetic factors that may inform future studies on everything from human population adaptations to skin pigmentation biology and related skin conditions such as melanoma.

The University of Pennsylvania's Sarah Tishkoff led a team of investigators who searched for genetic factors with ties to skin pigmentation genetics in nearly 1,600 genotyped participants from Africa with quantified skin tones. Along with extensive skin tone variation across the populations considered, they saw signs that both light and dark pigmentation have likely been evolving in the human lineage for tens or hundreds of thousands of years.

The team's findings, published online today in Science, also highlighted the power of identifying previously unappreciated contributors to complex traits by looking beyond more commonly studied European populations.

"[Prior] studies never found some of the genes and the variants we found because — probably due to natural selection — the variants associated with light skin swept to very high frequency in Eurasians, particularly in Europe," Tishkoff said. "So, if they're at almost 100 percent frequency, there's no power to detect them in a genome-wide association study."

So far, skin pigmentation genetics has primarily been studied in European populations, she explained, leading to a potentially biased and incomplete view of the trait. And though her team observed large variations in skin pigmentation while working on other studies in Africa, "we never knew what was the cause of it."

"I've always been curious about why the San are so light-skinned relative to other Africans? And why are populations from East Africa that come from southern Sudan so dark-skinned?" Tishkoff said. "My lab is interested in trying to understand the genetic basis of adaptation and human evolutionary history, and skin color is one of the most strikingly variable traits that's thought to be adaptive."

For their new analysis, she and her colleagues genotyped 1,570 individuals from Ethiopia, Tanzania, and Botswana, focusing on millions of SNPs imputed or directly evaluated with Illumina Infinium Omni5 arrays. To gauge skin pigment variation in these African populations, they also measured and quantified light reflectance at inner arm skin sites where sun exposure should be relatively low.

The team then used 1000 Genomes Project sequences and whole genomes representing 135 Africans for further imputation and fine mapping, particularly around potentially causal variants. The most significant ties to skin pigmentation turned up in the chromosome 15 gene SLC24A5, which was previously implicated in pigmentation in European populations.

A potentially causal non-synonymous variant at that site is common in populations from Europe, Pakistan, and India, the researchers discovered when they tapped into global sequence data from the Simons Genome Diversity Project. But it also turned up with frequencies ranging from 28 to 50 percent in some populations in Ethiopia and Tanzania.

From these, and other clues, they concluded that migration and gene flow from outside of Africa brought that light pigment-linked variant to East Africa, where it continued to spread and was subjected to selection.

"The variant in Africa is on the same haplotype background, so it had to have arisen, we think, by migration from out of Africa back into Africa probably … at least 5,000 years ago," Tishkoff said. "This was definitely pre-colonial."

New skin pigmentation associations turned up, too. For example, the team identified an ancestral light pigment-associated haplotype in and around a largely uncharacterized gene called MFSD12. A derived MFSD12 haplotype linked to darker skin appeared to be shared between most populations in sub-Saharan Africa as well as populations in South Asia and Australo-Melanesia.

When they used mouse or zebrafish gene knockdown experiments to take a closer look at that locus, the researchers found that the ancestral, light pigment-related haplotype led to enhanced MFSD12 expression in lysosome-like structures in melanocyte cells. That, in turn, bumped up production of a yellow-tinged pigment, investigators found, while curbing production of darker eumelanin compounds. Conversely, dialing down MFSD12 expression converted rust-colored agouti mice to a medium grey hue.

At the DDB1/TMEM138 locus, meanwhile, the team saw overlap with sequences already implicated in ultraviolet light response, melanoma, and/or DNA repair. In this linkage disequilibrium-rich region, the group detected derived light and dark pigmentation alleles stretching back some 250,000 years and 600,000 years, respectively, in the human lineage.

One of these, a dark pigment-associated allele neighboring TMEM138 turned up in populations in Nilo-Saharan populations in East Africa, along with a significant subset of individuals from populations in South Asia and Australo-Melanesia. On the other hand, the team estimated that a light pigment-related haplotype in this region seems to have swept to fixation in Eurasians some time within the last 60,000 years or so — since modern humans migrated out of Africa.

The association study also unearthed new variants at a OCA2/HERC2 locus already implicated in pigmentation, consistent with potential convergent evolution in this region. Digging into this region, the researchers uncovered an alternative splice variant in an albinism-related region of OCA2 that leads to a truncated transcript over-represented in individuals with lighter skin.

Based on their experiments, they concluded that it's likely the long-to-short OCA2 transcript ratio that mediates pigmentation. "People who have the dark allele have much more of the long transcript," she said, "and people who have two copies of the light allele have … more of the shorter transcript." 

The researchers are continuing to explore the potential functions and pleiotropic effects of the pigmentation-associated variants identified so far. The distribution of skin pigment haplotypes may also provide new insights into human population histories and relationships, Tishkoff noted, informing everything from out of Africa migration models to our perception of the potential skin tones present in our shared human ancestors.