Combined use of feature engineering and machine-learning to predict essential genes in Drosophila melanogaster.
Combined use of feature engineering and machine-learning to predict essential genes in Drosophila melanogaster.
复制标题
DOI：
10.1093/nargab/lqaa051
复制
发表时间：
2020-09
影响因子：
4.6
通讯作者：
Young ND
中科院分区：
文献类型：
--
作者：
Campos TL;Korhonen PK;Hofmann A;Gasser RB;Young ND
表征对生物体生存至关重要（即必需）的基因对于深入了解维持生命的基本细胞和分子机制非常重要。对醋蝇（果蝇）的功能基因组研究已经揭示了该模型物种的许多基因的功能，但表型实验的结果有时可能不明确。此外，人们对基因重要性的特征知之甚少，这给计算预测带来了挑战。在这里，我们利用公开的黑腹果蝇全面的基因组表型数据集和基于机器学习的工作流程来预测这种果蝇的基本基因。我们发现了此类基因的强大预测因子，为计算预测较少研究的节肢动物害虫和传染病媒介的重要性铺平了道路。
Characterizing genes that are critical for the survival of an organism (i.e. essential) is important to gain a deep understanding of the fundamental cellular and molecular mechanisms that sustain life. Functional genomic investigations of the vinegar fly, Drosophila melanogaster, have unravelled the functions of numerous genes of this model species, but results from phenomic experiments can sometimes be ambiguous. Moreover, the features underlying gene essentiality are poorly understood, posing challenges for computational prediction. Here, we harnessed comprehensive genomic-phenomic datasets publicly available for D. melanogaster and a machine-learning-based workflow to predict essential genes of this fly. We discovered strong predictors of such genes, paving the way for computational predictions of essentiality in less-studied arthropod pests and vectors of infectious diseases.