(NewsNation) — Dating someone who might be your doppelganger is all fun and games until you realize they might be your relative.

This fact may have to be in the back of the minds of mingling singles, as new research in the Aug. 23 edition of the Cell Reports journal shows a strong correlation between facial similarities and genetic variants.

“Our study provides a rare insight into human likeness by showing that people with extreme look-alike faces share common genotypes, whereas they are discordant at the epigenome and microbiome levels,” senior author Manel Esteller, of the Josep Carreras Leukaemia Research Institute in Barcelona, Spain, said in a statement. “Genomics clusters them together, and the rest sets them apart.”

Esteller and his team were able to reach this conclusion by using the work of François Brunelle — a Canadian photographer who has been collecting pictures of look-alikes worldwide since 1999.

Using these pictures, they narrowed the group of 32 look-alike pairs to the closest-looking pair of 16 based on three different facial recognition algorithms. The new group was asked to fill out biometric and lifestyle questionnaires and produce saliva DNA for multi-omics analysis.

According to Esteller, these sets of tests reveal to what extent these biological factors — from our blood to our guts and how they all work together — go into how we look.

“This unique set of samples has allowed us to study how genomics, epigenomics and microbiomics can contribute to human resemblance,” Esteller explained to EurekAlert!.

After narrowing the research pool to 16 couples as test subjects, it was determined that they shared similar genotypes overall, even though they do not share the same DNA sequence.

Furthermore, it was concluded that nine out of the 16 pairs whom artificial intelligence determined bore the closest resemblance had 19,277 genetic variations in common that included shared physical behavior and traits, such as height, weight and smoking habits.

To put this into context, individuals rarely have identical characteristics. The most common variation amongst people occurs once in every 300 bases of DNA, of which we all share the same four instructions (ACGT).

Because it’s hard to get a hold of “look-alike data,” as the study describes it, the sample size ended up being small, resulting in shortcomings such as skewed results. For example, any results from non-genetic data could relate to an underpowered study. 

Furthermore, because the pictures are two dimensional and in black and white, the study admits that “valuable” information, such as “subtle skin tones, and unique facial features” are lacking.