中国科学院昆明动物研究所机构知识库

明晰适应性表型变异的分子基础是理解生物多样性形成与维持的核心。根据Gommans等的理论，遗传变异分子机制如果能提高生物体产生新表型以适应环境变化的能力，该机制便可能会增加物种的可进化性，如此，高遗传变异水平或表型可塑性的体系往往会更易被选择。而相较于基因组变异“全或无”的特性，转录和转录后水平修饰的多样度和灵活性更高，它们可能正是真核生物的可进化能力显著强于原核生物的重要体现。家鸡自其野生祖先--红原鸡（red junglefowl, RJF）被驯化后，演化出相当丰富的品种，是表型多样度最高的鸟类。但鸡基因组很小（不到1G），要在仅仅8,000年的分歧时间里产生如此高的表型多样度，若只依靠基因组变异，显然是不够的。高通量测序技术的迅速发展，显著降低了分析的成本和周期。故此，本论文基于大范围基因组和转录组数据，构建鸡基因间长链非编码RNA（long intergenic non-coding RNA, lincRNA）和RNA编辑谱，分析它们的组织特性、功能以及其在家鸡驯化过程中的作用，并探讨了云南特有品种鸡--瓢鸡的无尾性状的分子遗传机制。基于821个RNA-Seq（RNA Sequencing）样本，我们鉴定得到了4,754个lincRNA基因，并分析其诸如表达、组织特异性以及转录本长度和外显子数等特征，发现lincRNA的表达水平要比蛋白编码基因低得多，但组织特异性更高，转录本短而外显子数少。通过比较转录组分析，我们找到了68个在不同条件下差异表达的lincRNA。而利用已有的全基因组关联研究（genome-wide association studies, GWAS）数据，我们发现47个与特定表型显著相关的lincRNA。我们的家鸡与红原鸡间的大范围群体基因组比较分析结果揭示，419个lincRNA可能在家鸡驯化过程中受到强烈的人工选择压力，它们或许通过影响鸡的行为和繁殖力等来参与新品种的形成。利用96个鸡RNA-Seq样本，我们构建了鸡的RNA编辑谱。分析不同组织的整体编辑水平模式发现，鸡里，肝脏中RNA编辑水平最低，骨骼肌中最高，这与哺乳动物肌肉中编辑水平相对较低的结论迥然相异。家鸡和红原鸡间RNA编辑水平比较分析揭示，两个群体间具有非常相似的编辑模式，群体间差异编辑的位点不到500个，仅约占所有A-to-I编辑位点的1%。两个群体的基因组比较分析结果亦表明，编辑位点在家鸡驯化过程中受到的选择压力很弱。而PhastCons值显示RNA编辑位点在鸟类中是保守的，但在其它脊椎动物中发生了快速进化。ADAR酶的RNA编辑贡献性分析也得到了与哺乳动物截然不同的现象：鸡里，ADARB1和ADARB2基因的表达与编辑水平呈显著正相关，而ADAR基因与编辑水平呈显著负相关；哺乳动物中，ADAR1和ADAR2酶是RNA编辑的主要贡献者，而ADAR3是RNA编辑的主要抑制子。不过，与哺乳动物中相一致的是，我们在鸡中也找到许多除ADAR酶以外的RNA编辑调节基因。例如，编码小亚基加工体的非核糖体蛋白基因UTP18可能通过影响核糖体的组装来抑制RNA编辑。基于21个不同胚胎发育时期的尾部转录组的比较分析，我们发现了许多与鸡尾部发育显著相关的差异表达基因和共表达模块，例如，MYL3和MYH7B参与的肌发生，或PAX9和PAX1涉及的脊柱形成。通过对20个成年瓢鸡与98个成年正常有尾鸡的共118个基因组重测序数据的群体基因组比较分析，我们找到两个在瓢鸡和正常有尾鸡间呈现显著遗传分化的区域，chr2: 84.5Mb - 89Mb和chr24: 6.12Mb - 6.18Mb。它们包含了诸如TERT，PDCD6等许多与骨发生、凋亡及钙信号通路等相关的基因。而尾部胚胎发育显微观察结果显示，瓢鸡胚胎尾部缺陷在胚胎发育第四天前就已形成。因此，我们推断，调节元件所受选择压力可能导致这些区域基因在胚胎发育早期时的活性改变，并通过促凋亡、BMP或者NF-κB等信号通路破坏尾牙的间质维持，最终导致不正常的尾部发育。本论文通过整合大范围转录组和基因组数据分析，揭示了转录及转录后水平修饰在家鸡驯化过程中的重要作用，完善了现有lincRNA和RNA编辑谱及其功能注释，并探讨了瓢鸡无尾性状发生的可能遗传机制。我们的工作将为今后复杂表型适应性进化机制的相关研究提供思路和奠定基础。 

Understanding molecular underpinnings of adaptive phenotypic alteration is the core of researches on biodiversity formation and maintenance. As Gommans et al. hypothesize, molecualr mechanisms underlying genetic variance would enhance specices evolvability if they raise organisms’ production capacities of new phenotypes to accommodate environmental changes. Therefore, systems with high genetic variation or phenotypic plasticity are more likely to be selected. Contrast to the ‘all or nothing’ property of genomic variation, transcriptional and post-transcriptional modification are more diversified and flexible, which are probably important manifestations of stronger evolvability of eukaryote than prokaryote. After being domesticated from their ancestor red junglefowl (RJF), chicken evolved into considerably plentiful breeds, and own the highest phenotypic diversity in birds. However, considering a small chicken genome (no more than 1Gb) and short divergence time (approximately 8,000 years) between domesticated chickens and RJFs, it is obviously not enough that such high phenotypic diversity only comes from genomic variation. Meanwhile, with the rapid advance of high-throughput sequencing, the analyzing cost and time have been significantly decreased. Hence, based on large-scale transcriptome and genome sequencing data, the thesis focuses on construction of long intergenic non-coding RNA (lincRNA) and RNA editing repertoires, analyses of their tissue features, function, and roles during chicken domestication, as well as study of genetic mechanisms underlying the rumpless trait of Piao chickens. Using 821 RNA Sequencing (RNA-Seq) samples, we totally identified 4,754 lincRNAs, and analyzed their features of expression, tissue specificity, as well as the length and exon number of transcripts. Compared to protein-coding genes, lincRNAs exhibited lower expression levels but higher tissue specificity, and had shorter transcripts with less exons. Through comparative transcriptomic analyses, we identified 68 lincRNAs that were differentially expressed under different conditions. By retrieving available genome-wide association studies (GWAS) data, we found 47 lincRNAs linked to special phenotypes. Our large-scale population genomic analysis reveals that 419 lincRNAs under artificial selection were potentially related to chicken domestication. These strongly selected lincRNAs may be involved in the origin of new breeds through influencing chicken behavior or fecundity.Based on 96 RNA-Seq samples, we constructed chicken RNA-editing repertoires. By studying overall editing patterns across different tissues, we observed that editing level was lowest in liver, and highest in skeletal muscle. The result is utterly different from mammals where the lowest editing level is found in skeletal muscle. Comparisons of RNA editing levels reveal that domesticated chickens and RJFs had similar editing patterns, and less than 500 differentially edited sites (about 1% of all A-to-I editing sites). Comparative genomic analysis between the two populations indicates poor selective pressures on RNA editing sites. PhastCons scores suggest that RNA editing sites were conserved in birds but evolved quickly in other vertebrates. Attribution assessment of ADAR enzymes to RNA editing displays a phenomenon distinct from mammals: in chicken, the expression levels of ADARB1 and ADARB2 were significantly correlated to editing levels, while ADAR were expressed in positive correlation to editing levels; in mammals, ADAR1 and ADAR2 enzymes are primary editors, whereas ADAR3 predominantly serves as an editing suppressor. Nevertheless, consistent with mammals, we detected many non-ADAR editing regulators in chicken. For example, UTP18, encoding a small-subunit (SSU) processome nonribosomal protein, may negatively regualte RNA editing through affecting eukaryotic ribosome assembly.By performing comparative analysis on 21 transcriptomes representing different tail development stages, we identified several differentially expressed genes and significant gene co-expression modules involved in chicken tail development, like MYL3 and MYH7B participating in myogenesis, as well as PAX9 and PAX1 refering to vertebral column formation. Comparative population genomic analysis on 118 resequencing genomes consisting of 20 Piao genomes and 98 controls with a normal tail, we found two regions showing significant genetic differentiation between Piao and normal-tailed chickens, i.e. chr2: 84.5Mb - 89Mb and chr24: 6.12Mb - 6.18Mb. The two regions include many genes like TERT and PDCD6 that are associated with skeletogenesis, apoptosis, or calcium signaling pathway. Embryos microscopic observation showed that the rumpless phenotype of Piao was established before the fourth day of embryo development. Thus, we hypothesize that selection pressures on the regulatory elements of some of these genes produced ectopic expressions during early embryo development that possibly destroy the mesenchymal maintenance in the tail bud through the proapoptotic, BMP, or NF-κB signaling pathways, and eventually result in failure of normal tail development. By combining large-scale transcriptomic and genomic analyses, the thesis demonstrates important roles of transcriptional and post-transcriptional modification during chicken domestication, improves the existing spectrum and functional annotations of lincRNA and RNA editing, and investigates mechanisms underlying the developmental retardation of Piao chicken tail. Our work provides insight into adaptive evolution of complex traits.