Can genomics wizards make an informed guesstimate on what a person looks like, based on his or her DNA? J. Craig Venter sure thinks so, per a new PNAS paper. Yet his Human Longevity Institute study is facing Twitter blowback for that claim.
Venter’s proof-of-concept study used a machine learning algorithm to analyze the genomic and biometric data of 1,061 volunteers. It looked at gender, facial structure, age, height, weight, skin color, eye color, and voice, generating a facsimile of the person based on their genetic analysis.
Former employee Jason Piper is listed as an author on the study, but he takes exception with its conclusion — and maintains that the methodology simply generates images that look like generic versions of any person’s given race and gender. For instance, the Venter algorithm only worked half the time when the researchers tried to work backward and match a face to the correct DNA sample from collection of genetic profiles of people with the same ethnic background. But it worked 8 out of 10 times when they tried to match a face to a DNA sample when the genetic profiles were drawn from people of an array of ethnic backgrounds.
About the Author Reprints
To submit a correction request, please visit our Contact Us page.
STAT encourages you to share your voice. We welcome your commentary, criticism, and expertise on our subscriber-only platform, STAT+ Connect