Algorithm predicts sexual orientation with 70 percent accuracy, using DNA.
According to a release from the American Society of Human Genetics, a newly developed algorithm using epigenetic information from nine regions of the human genone can predict the sexual orientation of males with a 70 percent accuracy rate.
Tuck C. Ngun, PhD, first author on the study and a postdoctoral researcher at the David Geffen School of Medicine of the University of California, Los Angeles said in the release, “To our knowledge, this is the first example of a predictive model for sexual orientation based on molecular markers.”
Going beyond the normal genetic information that is contained in the DNA, the scientists looked at the patterns of DNA mythylation across the genome in male identical twins. Methylation is a molecular modification to DNA that affects how strongly the gene is expressed. Environmental factors can cause DNA to be methylated differently, despite twins having the exact same genetic sequence.
The use of twins in the study enabled researchers to control the genetic differences while exposing the effect of the methylation. The study included 37 pairs of twins in which one of the twins was a homosexual, and 10 pairs in which both twins identified as homosexual.
Dr. Ngun and his team developed a machine learning algorithm named FuzzyForest, to help them analyze the more than 400,000 points in the data set. Using just nine regions at different points across the genome, the algorithm correctly identified the sexual orientation with an accuracy rate of 70 percent.
Dr. Ngun commented, “A challenge was that because we studied twins, their DNA methylation patterns were highly correlated. The high correlation and large data set made it difficult to identify differences between twins, determine which ones were relevant to sexual orientation, and determine which of those could be used predictively.”