Enabling Studies of Genome-Scale Regulatory Network Evolution in Large Phylogenies with MRTLE.
Enabling Studies of Genome-Scale Regulatory Network Evolution in Large Phylogenies with MRTLE.
复制标题
通过Mrtle的大系统发育中的基因组规模调节网络进化的研究。
DOI：
10.1007/978-1-0716-2257-5_24
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
中科院分区：
文献类型：
--
作者：
转录调控网络指定特定环境的基因模式，并在物种如何进化和适应中发挥核心作用。推断非模式物种的基因组尺度调控网络是研究调控网络的保护和分化模式的第一步。多种物种在不同环境刺激下获得的转录组学数据正变得越来越可用，这些数据可用于推断调控网络。然而，在系统发育背景下的多基因调控网络的推断和分析仍然具有挑战性。我们开发了一种算法，多物种调节网络学习（MRTLE），以促进这种调节网络进化的研究。MRTLE是一种基于概率图形模型的算法，它使用系统发育结构、多个物种的转录组数据和每个物种的序列特异性基序来同时推断多个物种的基因组尺度调控网络。我们利用在不同胁迫条件下收集的转录组学测量数据，应用MRTLE研究了六种子囊菌酵母的调控网络进化。MRTLE网络概括了模式生物酿酒酵母和非模式物种中实验衍生的相互作用，与不使用系统发育信息的方法相比，它更有利于网络推断。我们研究了跨物种的调控网络，发现与显著表达和网络变化相关的调控网络参与了与压力相关的过程。MTRLE及其相关的下游分析提供了一个可扩展的原则性框架，用于研究大系统发育中多个物种的转录调控网络的进化动力学。
Transcriptional regulatory networks specify context-specific patterns of genes and play a central role in how species evolve and adapt. Inferring genome-scale regulatory networks in non-model species is a first step for examining patterns of conservation and divergence of regulatory networks. Transcriptomic data obtained under varying environmental stimuli in multiple species are becoming increasingly available, which can be used to infer regulatory networks. However, inference and analysis of multiple gene regulatory networks in a phylogenetic setting remains challenging. We developed an algorithm, Multi-species Regulatory neTwork LEarning (MRTLE) to facilitate such studies of regulatory network evolution. MRTLE is a probabilistic graphical model-based algorithm that uses phylogenetic structure, transcriptomic data for multiple species, and sequence-specific motifs in each species to simultaneously infer genome-scale regulatory networks across multiple species. We applied MRTLE to study regulatory network evolution across six ascomycete yeasts using transcriptomic measurements collected across different stress conditions. MRTLE networks recapitulated experimentally derived interactions in the model organism S. cerevisiae as well as non-model species and it was more beneficial for network inference than methods that do not use phylogenetic information. We examined the regulatory networks across species and found that regulators associated with significant expression and network changes are involved in stress related processes. MTRLE and its associated downstream analysis provides a scalable and principled framework to examine evolutionary dynamics of transcriptional regulatory networks across multiple species in a large phylogeny.