Eye colour is commonly and broadly categorized as blue, brown, or something in between, like green, hazel, and gray. Sometimes, qualifiers like “light” or “-ish” can narrow the description, but subtle variations are rarely quantified.

Now a research team from the Netherlands has discovered three genetic loci that are associated with variations along the spectrum from the lightest blue to the darkest brown. Their study appears in PLoS Genetics today.

Led by Manfred Kayser from Erasmus University Medical Center, in Rotterdam, they looked at high-resolution photographs of 5,951 subjects and digitally quantified their eye colours into hue (H) and saturation (S) values. In this colour palette, S ranges between 0 and 1, and H ranges between 0 and 60. Kayser has blue eyes and, though he didn’t include himself in the study, he thinks his colour falls in the upper left corner, with H 50 and S 0.1.

His team conducted a genome-wide association study of the subjects and found three new loci associated with subtle variations of eye colour. Seven genes have been previously identified with categorical colours of the old blue-brown-intermediate model. And together with the three new loci, about 50% of eye colour variation can now be explained, according to a model the team used to predict eye colour.

The authors suggest this improved accuracy in prediction could assist in identifying unknown suspects in future forensic applications. Say you have a DNA profile from material left at a crime scene. You’ll want to match that with the suspects identified by the police or with some DNA database (like the FBI’s CODIS). “But if you cannot find a match, what do you do?” Kayser asks. His team is trying to develop a way to predict the appearance of the person-of-interest. By predicting eye colour using crime scene samples, police investigations can concentrate on a specific colour, rather than a vague and subjective category.

Image: Erasmus MC Rotterdam