Evidence · Replicated

What the research found

In the cited research, the T allele at rs12203592 was associated with darker natural hair colour, less often blonde.

CC: Lighter range
CT: Middle
TT: Darker range

Clear effect

What the study found

In the study, how many people said their natural hair colour as a young adult was blonde, by genotype:

CCNo copy: 1,139 of 6,817 (17%)
CTOne copy: 125 of 2,881 (4%)
TTTwo copies: 14 of 382 (4%)

US women and men and Australian adults of European ancestry (initial GWAS and four follow-up studies, pooled) · Table 5 (IRF4 RS12203592 wt, het and var columns read as CC, CT and TT; totals summed over the black, dark brown, light brown, blonde and red rows)

Evidence: Replicated genome-wide association
Who was studied: Primarily European ancestry

1000 Genomes Project

How common each genotype is

How common each genotype is in the 1000 Genomes Project, by region
Region	CCLighter range	CTMiddle	TTDarker range
All regions2,504 people	93%	6%	<1%
Africa661 people	98%	1%	<1%
The Americas347 people	86%	13%	<1%
East Asia504 people	All	Not seen	Not seen
Europe503 people	79%	20%	2%
South Asia489 people	99%	1%	Not seen

Among the 2,504 people of the 1000 Genomes Project, in five world regions. These percentages describe groups of people, never anyone’s chance of having a trait.

Ancient genomes

Through time

How often the studied letter was read in ancient people, era by era, in the region chosen.

Region

Through time: Europe
Before farmingbefore 6500 BC	46%26 of 56 copies read
First farmers6500 to 3000 BC	13%114 of 868 copies read
Bronze Age3000 to 1200 BC	14%100 of 728 copies read
Iron Age and Antiquity1200 BC to AD 600	12%124 of 1,018 copies read
Middle AgesAD 600 to 1500	7%36 of 552 copies read
Today1000 Genomes	12%117 of 1,006 copies

Neanderthal and Denisovan genomes

Altai Neanderthal: No copies
Vindija Neanderthal: No copies
Chagyrskaya Neanderthal: No copies
Denisovan: No copies

Neanderthals and Denisovans carried the other letter, so this letter is younger than our split with them.

For most ancient people only one of their two copies can be read, so these counts are copies read, not people. Close relatives, repeated samples and samples the archive marks as unreliable are left out. Damage in old DNA can fake this letter, so it is counted only from samples treated to remove that damage and from high-quality genomes. Ancient genomes from the Allen Ancient DNA Resource (AADR v66): 15,136 people.

rs12203592

The marker

Gene or region: IRF4
Marker: rs12203592
Location: Chromosome 6 · GRCh37 396,321 · GRCh38 396,321
Letters studied: C and T, with T as the studied allele

A DNA Traits report reads this marker in your raw DNA file and shows which of the three genotypes is yours, with how common it is. It counts the copies of the studied allele; it does not measure the trait itself.

Marker: IRF4 · rs12203592
Location: Chromosome 6 · GRCh37 396,321 · GRCh38 396,321
Studied allele: T
Published statistic: Per T on black (1) to blonde (4): -0.36 initial GWAS (P 6.6e-36), -0.35 pooled follow-up studies without the GWAS (P 2.0e-106), -0.35 all pooled (P 1.5e-137) (Han Table 4); 7.0% of residual variation (Results). Zhang 2013 Table 6: +0.35 per T toward black (P 1.0e-28). Hysi 2018 Supplementary Table 2: +0.375 SD per T toward black in UK Biobank and in 23andMe, VisiGen z 29.97. In Table 5 black hair rises from 3.8% (CC) to 16.5% (TT) (Table 5 (IRF4 RS12203592 wt, het and var columns read as CC, CT and TT; totals summed over the black, dark brown, light brown, blonde and red rows))
1000 Genomes Project phase 3, people with 0, 1 and 2 copies: All regions 2332 / 160 / 12 · Africa 651 / 9 / 1 · The Americas 299 / 46 / 2 · East Asia 504 / 0 / 0 · Europe 395 / 99 / 9 · South Asia 483 / 6 / 0

See this card in the sample report

Limits

What it cannot tell you

This is the freckling and hair greying cards' marker, so it is not independent evidence. One of many markers that shape hair colour.

A genome-wide association study identifies novel alleles associated with hair color and skin pigmentation (Han et al., 2008)

The same topic

More in Appearance

Free, once per account

See your own two letters.

Upload the raw DNA file you already have. Your report reads this marker and every other card studied for you, each with its evidence and its limits.

See a sample report