Researchers able to model facial features from DNA data

Image Credit: PNAS

Researchers from Human Longevity, Inc., have used machine learning to link DNA genome data to facial and physical traits and are now able to use facial modeling and other physical trait predictions to identity people. 

In a study with over 1,000 ethnically diverse participants ranging in age from 18 to 82, researchers were able to correctly identify about 80% of ethnically mixed participants and 50% of African American or European participants.

DNA genome information is used to create a facial image predicting face shape and features, including eye and skin color. It also is used predict sex, age, height and weight. Eye color, skin color and sex were predicted with the greatest accuracy. 

Researchers would like to expand their study to hundreds of thousands of participants to refine their algorithms, but are also highlighting the need for better safeguards of participants’ privacy. While most would probably agree that using new facial imaging technology to create realistic mugshots of criminals is a good thing, privacy issues and possible future uses need consideration.

If your DNA can predict your personality traits, could this information potentially be tapped into by possible employers? If your possible health problems can be predicted, would health insurance companies have access to this information and be able to use it against you? The research team advocates, “more public deliberation is needed as more and more genomes are generated and placed in public databases.”



Researchers from Human Longevity, Inc. Use Whole Genome Sequence Data and Machine Learning to Identify Individuals Through Face and Other Physical Trait Prediction

Human Longevity, Inc. Press Release, Sept. 5, 2017


Identification of individuals by trait prediction using whole-genome sequencing data

Authors Info

Article published on pnas.org, June 28, 2017