Cosplicing network analysis of mammalian brain RNA-Seq data using WGCNA and Mantel correlations

Gene network analysis aims to understand how large groups of genes coordinate their activity to support complex biological functions. Gene coexpression networks are constructed on the basis of coordinated change in expression levels. However, a majority of genes have multiple versions or isoforms, often with distinct functionality. Gene cosplicing networks aim to capture coordinated changes in either isoform levels or relative ratios for different genes, which in principle could happen even with constant overall expression levels. We developed a method of capturing coordinated changes in isoform production, based on correlated exon inclusion rates. Integrating these correlations transcriptome wide results in a cosplicing network. The genes deemed important in the cosplicing network (hubs) include long many-exon genes with essential neurological functionality and association with autism spectrum disorders. Importantly, the cosplicing patterns are found, albeit with alterations, across species (mouse, macaque and human) and brain regions (striatum and cortex). 60 samples of HSCC animals tested for catalepsy were analyzed: 28 females (13 non-responders, 15 responders), 32 males (17 non-responders, 15 responders).基因网络分析旨在探究大规模基因群体如何协调其表达活性，以支撑复杂的生物学功能。基因共表达网络（Gene coexpression networks）构建于基因表达水平的协同变化基础之上。然而，大多数基因存在多种变体或转录本异构体（isoforms），且通常具备不同的生物学功能。共剪接网络（Gene cosplicing networks）旨在捕捉不同基因的转录本异构体水平或相对比例的协同变化，这类变化在总表达水平保持恒定的情况下理论上仍可发生。本研究基于外显子包含率的相关性，开发了一种捕捉转录本异构体生成协同变化的方法。将这些全转录组范围内的相关性进行整合，即可构建得到共剪接网络。在共剪接网络中被认定为关键的基因（即枢纽基因，hubs），包含具备核心神经生物学功能且与孤独症谱系障碍相关的长链多外显子基因。值得注意的是，尽管存在一定程度的差异，共剪接模式在多个物种（小鼠、猕猴与人类）以及多个脑区（纹状体与大脑皮层）中均被观测到。本研究对60只经僵直症（catalepsy）检测的HSCC模型动物样本进行了分析：其中雌性样本28只（无响应者13只，响应者15只），雄性样本32只（无响应者17只，响应者15只）。