Gene Regulatory Network Inference using 3D Convolutional Neural Network

Gene regulatory networks (GRNs) consist of gene regulations between transcription factors (TFs) and their target genes. Single-cell RNA sequencing (scRNA-seq) brings both opportunities challenges to the inference GRNs. On one hand, scRNA-seq data reveals statistic information expressions at single-cell resolution, which is conducive construction GRNs; on other noises dropouts pose great difficulties analysis data, causing low prediction accuracy by traditional methods. In this paper, we propose 3D Co-Expression Matrix Analysis (3DCEMA), predicts relationships classifying co-expression matrices triples using a convolutional neural network. We found that introducing third as comparison factor, our method can avoid disturbance dropouts, significantly increase pairs. Compared with existing GRN algorithms in-silico datasets scRNA-Seq datasets, algorithm based deep learning shows higher stability in task inference.