Using Deep Learning to Understand Genetic Networks

Carnegie Mellon University computer scientists have taken a deep learning method that has revolutionized face recognition and other image-based applications in recent years and redirected its power to explore the relationships between genes. The trick, they say, is to transform massive amounts of gene expression data into something more image-like. Then convolutional neural networks (CNNs), which are adept at analyzing visual imagery, can infer how the genes interact. Their results were published today in PNAS.

“CNNs, which were developed a decade ago, are revolutionary," says senior author Ziv Bar-Joseph. “I’m still in awe of Google Photos, which uses them for facial recognition. We sometimes take this technology for granted because we use it all the time. But it’s incredibly powerful and is not restricted to images. It’s all a matter of how you represent your data.”

In this case, he and Yuan were looking at gene relationships. The approximately 20,000 genes in humans work in concert, so it’s necessary to know how genes work together in complexes or networks to understand human development or diseases.

One way to infer these relationships is to look at gene expression. Generally, if gene A is active at the same time gene B is active, that’s a clue that the two are interacting, says first author Ye Yuan. Still, it’s possible that this is a coincidence or that both are activated by a third gene C. Several previous methods have been developed to tease out these relationships.

To employ CNNs to help analyze gene relationships, Yuan and Bar-Joseph used single-cell expression data. The results of hundreds of thousands of these single-cell analyses were then arranged in the form of a matrix or histogram so that each cell of the matrix represented a different level of co-expression for a pair of genes.

Presenting the data in this way added a spatial aspect that made the data more image-like and thus more accessible to CNNs. By using data from genes whose interactions already had been established, the researchers were able to train the CNNs to recognize which genes were interacting and which weren’t based on the visual patterns in the data matrix.

“It’s very, very hard to distinguish between causality and correlation,” Yuan says, but the CNNC method proved statistically more accurate than existing methods. The scientists anticipate that CNNC will become a key tool to analyze large datasets.