Department of Computer Science, University of California, Davis, CA, USA.
Genome Center, University of California, Davis, CA, USA.
Bioinformatics. 2019 Jul 1;35(13):2226-2234. doi: 10.1093/bioinformatics/bty945.
Gene expression prediction is one of the grand challenges in computational biology. The availability of transcriptomics data combined with recent advances in artificial neural networks provide an unprecedented opportunity to create predictive models of gene expression with far reaching applications.
We present the Genetic Neural Network (GNN), an artificial neural network for predicting genome-wide gene expression given gene knockouts and master regulator perturbations. In its core, the GNN maps existing gene regulatory information in its architecture and it uses cell nodes that have been specifically designed to capture the dependencies and non-linear dynamics that exist in gene networks. These two key features make the GNN architecture capable to capture complex relationships without the need of large training datasets. As a result, GNNs were 40% more accurate on average than competing architectures (MLP, RNN, BiRNN) when compared on hundreds of curated and inferred transcription modules. Our results argue that GNNs can become the architecture of choice when building predictors of gene expression from exponentially growing corpus of genome-wide transcriptomics data.
Supplementary data are available at Bioinformatics online.
基因表达预测是计算生物学中的重大挑战之一。转录组数据的可用性以及人工神经网络的最新进展为创建具有广泛应用的基因表达预测模型提供了前所未有的机会。
我们提出了遗传神经网络（GNN），这是一种用于预测全基因组基因表达的人工神经网络，给定基因敲除和主调控因子扰动。在其核心，GNN 在其架构中映射现有的基因调控信息，并且它使用专门设计的细胞节点来捕获基因网络中存在的依赖关系和非线性动态。这两个关键特征使得 GNN 架构能够在不需要大型训练数据集的情况下捕捉复杂的关系。结果，与竞争架构（MLP、RNN、BiRNN）相比，GNN 在数百个经过精心整理和推断的转录模块上的平均准确率提高了 40%。我们的结果表明，当从指数增长的全基因组转录组学数据集中构建基因表达预测器时，GNN 可以成为首选架构。
补充数据可在 Bioinformatics 在线获得。
