大脑皮层是哺乳动物复杂认知能力的生物学基础，高度进化的大脑皮层使我们拥有远超其他物种的智慧特征和更高阶的感知、思维能力。与啮齿类哺乳动物相比，高级灵长类的大脑皮层显现出多个方面的独特性，多尺度的演化结果导致人类新皮层体积和表面积更大，同时神经细胞种类及突触连接更加多样和复杂。胚胎期大脑皮层的产生伴随着发育过程中一系列生物学事件，这种调控方式受到严格的进化限制，对新皮层的正确组装和发挥功能至关重要。基于高通量转录组学的研究方法加速了对发育状态下大脑结构和功能相关基因表达程序的理解。目前多以小鼠作为模式动物进行神经发育的研究，但由于小鼠和人类的大脑在解剖学、生理学、神经细胞类型等方面存在相当大的差异，导致我们对人类新皮层神经发生转录调控的认识仍然十分有限。恒河猴在进化层面与人类的亲缘关系更接近，二者的皮层发育具有更多的共同特征，包括神经干细胞的复杂组成和较厚的浅层皮层层次结构，也具有更多相似的基因表达模式。

为了描述人类代表的高级灵长类皮层发育过程中干细胞分化到神经元转录组的时序性变化，我们建立了恒河猴的胚胎发育模型，覆盖了胚胎期神经发生和皮层层次形成的关键时间点（E40为神经发生开始，E50为Layer6形成，E70为Layer5形成，E80为Layer4形成，E90为Layer3和Layer2形成），通过剖腹产的手段终止妊娠，并在恒河猴胎脑的顶叶背侧皮层取样进行单细胞转录组测序。

首先，利用获得了53295个细胞单细胞转录组数据，构建了胚胎期恒河猴新皮层顶叶的单细胞转录组图谱。证实恒河猴早期主要进行“神经发生（neurogenesis）”，发育后期开始“胶质发生（gliogenesis）”过程，遵循保守的皮层发育模式。鉴定了恒河猴兴奋性神经元亚型和神经前体细胞的多样性，证明了恒河猴深层的神经元在发育早期先产生，浅层的兴奋性神经元在发育晚期后产生。

为了探究恒河猴神经前体细胞不同分化轨迹的差异，通过拟时序分析方法对脑室外放射状胶质细胞（outer radial glia, oRG）和中间祖细胞（intermediate progenitor cell, IPC）的生成过程进行了分化轨迹推断，并探究了伴随拟时序分化过程基因表达强度的动态变化。

通过LIGER的分析方法，我们将本研究产生的恒河猴胚胎期皮层发育的单细胞数据与已发表研究中的一些小鼠和人类胚胎期皮层发育单细胞数据集进行了整合。并在数据中验证了高级灵长类有的神经干细胞oRG（标记基因HOPX、MOXD1、FAM107A和CLU），在人类和小鼠数据特异性存在。通过对三个物种的神经前体细胞进行拟时序轨迹推断，证实中间前体细胞的生成轨迹在三个物种中都是保守的， oRG的生成轨迹只在人类与恒河猴数据集存在。随后，通过比较物种间不同发育阶段深层/浅层兴奋性神经元的数量比例，验证了人与恒河猴所属的高等灵长类浅层神经元生发期在皮层发育阶段的时长占比，明显高于小鼠这种啮齿类模式动物。

通过对三个物种胚胎期皮层发育过程中脑室区放射状胶质细胞（ventricular radial glia, vRG）基因的表达进行时序性分析，首先筛选出在皮层发育期间表达有显著性差异的基因（DEGs），根据时序性的表达趋势将其分为5类，type1与type2都是在初期表达最强然后表达强度逐渐降低，type3与type4是表达先逐渐变强然后减弱，type5中的基因就是随着发育过程表达逐渐增强。我们着重关注了其中的转录因子和RNA结合蛋白，发现在72个转录因子中大多数转录因子的表达具有相似的时序性趋势，在三个物种的vRG细胞中都是非常保守的，如NEUROD1、NEUROG2、POU3F2、ETV1、ETV5和FOS基因。

此外，通过scenic对恒河猴和小鼠的vRG进行了转录因子调控网络分析，筛选出每个时间点调控活性最高的转录因子，并且将筛选到的转录因子与其调控的靶基因构建蛋白质互作网络，结果发现恒河猴与小鼠的调控网络并不完全相同，EOMES、NEUROD1和EMX1等经典发育转录因子在两个物种调控网络中的作用具有共同点，FOXG1在恒河猴vRG转录因子调节网络中处于较为重要的位置，推测FOXG1可能与灵长类新皮层的扩张相关，为后续的研究提供了线索。

       总之，通过对覆盖恒河猴胚胎期皮层层次形成过程的单细胞转录组测序和分析，我们首次从分子动力学层面时序性解析了恒河猴神经前体细胞分化的过程。研究结果为解析高等灵长类皮层发育的进化特征以及恒河猴作为非人灵长类模式动物在神经发育研究中的应用提供了稀有的数据资源。

The cerebral cortex is the biological basis of complex cognitive ability in mammals. The highly evolved cerebral cortex gives us far more intelligent features, higher-order perception, and thinking abilities than other species. Compared with rodent mammals, the cerebral cortex of higher primates shows multiple aspects of uniqueness. The multi-scale evolution results in a larger volume and surface area of the human neocortex and more diverse and complex types of nerve cells and synaptic connections. A series of biological events during development accompanies the generation of the embryonic cerebral cortex. Evolution strictly limits this regulation and is essential for the correct assembly and function of the neocortex. Research methods based on high-throughput transcriptomics have accelerated the understanding of gene expression programs related to brain structure and function in developmental states. Most of the research on neurodevelopment is based on mice as model animals. However, due to the considerable differences in anatomy, physiology, and nerve cell types between mouse and human brains, our understanding of the transcriptional regulation of human neocortex neurogenesis still needs to be improved. Macaque monkeys are more closely related to humans at the evolutionary level. The cortical development of the two has more standard features, including the complex composition of neural stem cells and a thicker shallow cortical hierarchy.

In order to describe the temporal changes of stem cell differentiation to neuronal transcriptome during the development of advanced primate cortex represented by humans, we established an embryonic development model of macaque monkeys, covering the time point of cortical layer formation during the whole stage of embryonic neurogenesis ( E40 is the beginning of neurogenesis, E50 is the formation of Layer6, E70 is the formation of Layer5, E80 is the formation of Layer4, E90 is the formation of Layer2 and Layer3 ). Pregnancy was terminated by cesarean section, and single-cell transcriptome sequencing was performed on the dorsal parietal cortex of the macaque fetal brain.

Firstly, a single-cell transcriptome map of the parietal lobe of the neocortex of embryonic macaque monkeys was constructed using 53295 cell single-cell transcriptome data. It was confirmed that macaque monkeys mainly performed ' neurogenesis ' in the early stage and began ' gliogenesis ' in the late stage of development, following a conservative cortical development model. The diversity of excitatory neuron subtypes in macaque monkeys was also found, which proved that the deep neurons of macaque monkeys were produced in the early stage of development, and the shallow excitatory neurons were produced in the late stage of development.

With the LIGER analysis method, we integrated the single-cell data of macaque monkey embryonic cortical development generated in this study with some single-cell data sets of mouse and human embryonic cortical development in published studies. In the data, it was verified that the advanced primate neural stem cells oRG (marker genes HOPX, MOXD1, FAM107A, and CLU) were specific in human and mouse data. By inferring the temporal trajectory of neural precursor cells in three species, it is confirmed that the generation trajectory of intermediate precursor cells is conservative in all three species. The generation trajectory of oRG is unique to the human and macaque monkey datasets. Subsequently, by comparing the proportion of deep/upper excitatory neurons at different developmental stages between species, it was verified that the proportion of the upper layer neurons of higher primates belonging to humans and macaque monkeys in the cortical development stage was significantly higher than that of mice, a rodent model animal.

By the temporal analysis of the expression of vRG (radial glial cells in the ventricle area) genes during the embryonic cortical development of the three species, the genes (DEGs ) with significant differences in expression during cortical development were first screened out. According to the temporal expression trend, they were divided into five categories. Type1 and type2 were the strongest in the early stage, and then the expression intensity gradually decreased. Type 3 and type 4 were gradually increased and then decreased. The genes in type5 were gradually increased with development. We focused on the transcription factors and RNA-binding proteins. The expression of most transcription factors in 72 transcription factors has a similar temporal trend and is very conserved in vRG cells of three species, such as NEUROD1, NEUROG2, POU3F2, ETV1, ETV5, and FOS genes. It is proved that the gene expression of vRG differentiation is relatively conservative among humans, macaque monkeys, and mice.

The transcription factor regulatory network of macaque monkey and mouse vRG (radial glial cells in the ventricle ) was analyzed using scenic. The transcription factors with the highest regulatory activity at each time point were screened. The screened transcription factors and their regulated target genes constructed a protein interaction network. The results showed that the regulatory networks of the macaque monkey and mouse were not precisely the same. The classical developmental transcription factors, such as EOMES, NEUROD1, and EMX1, had typical roles in the regulatory networks of the two species. FOXG1 is essential in the regulation network of vRG transcription factors in macaque monkeys. FOXG1 may be related to the expansion of the primate neocortex, which provides clues for subsequent research.

In summary, by sequencing and analyzing the single-cell transcriptome covering the formation process of macaque monkey embryonic cortical layers, we first analyzed the differentiation process of macaque monkey neural precursor cells from the molecular dynamics level. The results provide valuable data resources for analyzing the evolutionary characteristics of cortical development in higher primates and the application of macaque monkeys as non-human primate model animals in neurodevelopmental research species.