11 September 2017
ByAppeared in BioNews 917
Dr Venter and colleagues at the company he co-founded, Human Longevity, Inc. (HLI) in San Diego, used DNA to match eight out of ten faces from an ethnically-diverse group of photographs.
'We set out to do this study to prove that your genome codes for everything that makes you, you,' said Dr Venter.
He also raised privacy issues, adding: 'We are also concerned that the public and the research community at large are not adequately focused on the need for better safeguards and policies for individual privacy in the genomics era and are urging more analysis, better technical solutions, and continued discussion.'
However, the results were met with immediate scepticism from scientists on social media.
The researchers, led by Dr Christoph Lippert at HLI, sequenced the genomes of 1601 people. They also captured high resolution 3D images of the participants' faces, and recorded each person's age, height, weight, and eye and skin colour.
They developed a software algorithm to use the DNA sequence data to build predicted images which they could then compare with photographs. For eight out of ten people, there was a match. But this fell to five out of ten faces when the test was limited to a single race.
'The face prediction is just predicting the average face for your race. You will always say, "Wow, that kind of looks like me",' said Dr Jason Piper, a data engineer at Apple who formerly worked at HLI. 'Because everyone looks close to the average of their race, everyone looks like their prediction.'
Critics said that because the team used a person's race and sex to predict average human faces rather than specific ones, the approach is not especially new.
Dr Venter's team acknowledges that their algorithm needs more genomic data in order to become more precise in its predictions.
'Calling it predicting from the genome is what's wrong,' said Dr Mark Shriver, an anthropologist at Pennsylvania State University at University Park, Pennsylvania, and a researcher in genes-to-face predictions. He told MIT Technology Review: 'The main message is way overstated. They just didn’t have enough people to find the genes that distinguish people. This is not the paper that is going to convince people that this is going to affect privacy or help forensics.'
The study was published in PNAS.