Netotea Sergiu, Sundell David, Street Nathaniel R, Hvidsten Torgeir R
Umeå Plant Science Center, Department of Plant Physiology, Umeå University, Umeå, Sweden.
BMC Genomics. 2014 Feb 6;15:106. doi: 10.1186/1471-2164-15-106.
Divergence in gene regulation has emerged as a key mechanism underlying species differentiation. Comparative analysis of co-expression networks across species can reveal conservation and divergence in the regulation of genes.
We inferred co-expression networks of A. thaliana, Populus spp. and O. sativa using state-of-the-art methods based on mutual information and context likelihood of relatedness, and conducted a comprehensive comparison of these networks across a range of co-expression thresholds. In addition to quantifying gene-gene link and network neighbourhood conservation, we also applied recent advancements in network analysis to do cross-species comparisons of network properties such as scale free characteristics and gene centrality as well as network motifs. We found that in all species the networks emerged as scale free only above a certain co-expression threshold, and that the high-centrality genes upholding this organization tended to be conserved. Network motifs, in particular the feed-forward loop, were found to be significantly enriched in specific functional subnetworks but where much less conserved across species than gene centrality. Although individual gene-gene co-expression had massively diverged, up to ~80% of the genes still had a significantly conserved network neighbourhood. For genes with multiple predicted orthologs, about half had one ortholog with conserved regulation and another ortholog with diverged or non-conserved regulation. Furthermore, the most sequence similar ortholog was not the one with the most conserved gene regulation in over half of the cases.
We have provided a comprehensive analysis of gene regulation evolution in plants and built a web tool for Comparative analysis of Plant co-Expression networks (ComPlEx, http://complex.plantgenie.org/). The tool can be particularly useful for identifying the ortholog with the most conserved regulation among several sequence-similar alternatives and can thus be of practical importance in e.g. finding candidate genes for perturbation experiments.
基因调控的差异已成为物种分化的关键机制。跨物种共表达网络的比较分析可以揭示基因调控中的保守性和差异性。
我们使用基于互信息和相关性上下文似然性的先进方法推断了拟南芥、杨树和水稻的共表达网络，并在一系列共表达阈值范围内对这些网络进行了全面比较。除了量化基因-基因连接和网络邻域保守性外，我们还应用网络分析的最新进展对网络属性进行跨物种比较，如无标度特征、基因中心性以及网络基序。我们发现，在所有物种中，网络仅在高于某个共表达阈值时才呈现无标度特性，并且维持这种组织的高中心性基因往往是保守的。网络基序，尤其是前馈环，在特定功能子网中显著富集，但在物种间的保守性远低于基因中心性。尽管个体基因-基因共表达已发生大量分化，但高达约80%的基因仍具有显著保守的网络邻域。对于具有多个预测直系同源基因的基因，约一半的基因有一个调控保守的直系同源基因和另一个调控分化或不保守的直系同源基因。此外，在超过一半的情况下，序列最相似的直系同源基因并非调控最保守的基因。
我们对植物基因调控进化进行了全面分析，并构建了一个用于植物共表达网络比较分析的网络工具（ComPlEx，http://complex.plantgenie.org/）。该工具对于在多个序列相似的备选基因中识别调控最保守的直系同源基因特别有用，因此在例如寻找扰动实验的候选基因方面具有实际重要性。
