Kwon Soo Bin, Ernst Jason
Bioinformatics Interdepartmental Program, University of California, Los Angeles, CA, USA.
Department of Biological Chemistry, University of California, Los Angeles, CA, USA.
Nat Commun. 2021 May 3;12(1):2495. doi: 10.1038/s41467-021-22653-8.
Identifying genomic regions with functional genomic properties that are conserved between human and mouse is an important challenge in the context of mouse model studies. To address this, we develop a method to learn a score of evidence of conservation at the functional genomics level by integrating information from a compendium of epigenomic, transcription factor binding, and transcriptomic data from human and mouse. The method, Learning Evidence of Conservation from Integrated Functional genomic annotations (LECIF), trains neural networks to generate this score for the human and mouse genomes. The resulting LECIF score highlights human and mouse regions with shared functional genomic properties and captures correspondence of biologically similar human and mouse annotations. Analysis with independent datasets shows the score also highlights loci associated with similar phenotypes in both species. LECIF will be a resource for mouse model studies by identifying loci whose functional genomic properties are likely conserved.
在小鼠模型研究的背景下，识别在人类和小鼠之间具有保守功能基因组特性的基因组区域是一项重大挑战。为解决这一问题，我们开发了一种方法，通过整合来自人类和小鼠的表观基因组、转录因子结合及转录组数据汇编中的信息，在功能基因组水平上学习保守性证据得分。该方法，即从整合功能基因组注释中学习保守性证据（LECIF），训练神经网络为人类和小鼠基因组生成此得分。所得的LECIF得分突出了具有共享功能基因组特性的人类和小鼠区域，并捕捉了生物学上相似的人类和小鼠注释的对应关系。使用独立数据集进行的分析表明，该得分还突出了两个物种中与相似表型相关的基因座。LECIF将通过识别其功能基因组特性可能保守的基因座，为小鼠模型研究提供资源。
