10X单细胞、10X空间转录组和三代全长联合分析的文章（文献分享）

admin管理员组

文章数量:1622628

今天给大家分享一篇文献 ，文献的名称叫做A spatially resolved brain region- and cell typespecific isoform atlas of the postnatal mouse brain,文献的地址在这里,这篇文献于2021年发表于NC，其实的精髓在于运用了三种单细胞技术，10X单细胞、10X空间转录组和长片段分析，值得我们借鉴一下，我们今天的任务就是来分享一下这篇文献。intensive reading~~

（1）简介

Splicing varies across brain regions, but the single-cell resolution of regional variation is unclear.（脑区的可变剪切信息unclear），We present a single-cell investigation of differential isoform expression (DIE) between brain regions using single-cell long-read sequencing in mouse hippocampus and prefrontal cortex in 45 cell types at postnatal day 7（对海马和前额叶皮层 进行 single-cell long-read sequencing研究），Isoform tests
for DIE show better performance than exon tests.（研究效果比外显子好），We detect hundreds of DIE events traceable to cell types, often corresponding to functionally distinct protein isoforms（检测到数百种可溯源到细胞类型的DIE事件，这些事件通常对应于功能不同的蛋白质同工型 ），Mostly,one cell type is responsible for brain-region specific DIE.（一种细胞类型responsible for大脑具体区域的IDE事件），However, for fewer genes, multiple cell types influence DIE.（较少的基因，多种细胞类型影响DIE）因此，区域识别可以（尽管很少）override细胞类型的特异性。 Cell types indigenous to one anatomic structure display distinctive DIE（一种解剖结构固有的细胞类型显示出独特的DIE ），e.g. the choroid plexus epithelium manifests distinct transcription-start-site usage.（有点生涩，哈哈），Spatial transcriptomics and long-read sequencing yield a spatially resolved splicing map（联合分析）。Our methods quantify isoform expression with cell-type and spatial resolution and it contributes to further our understanding of how the brain integrates molecular and cellular complexity（意义）。
看来文章主要是研究可变剪切的空间区域性，以及细胞特异性

（2）Introduction

Alternative splicing (AS) affects almost all spliced genes in mammals，vastly expands the proteome and increases functional diversity of cell types. Alternative transcription start sites (TSS) and poly-adenylation (polyA) sites further expand the alternative isoform landscape, regulating development,
differentiation, and disease。（可变剪切的作用），These RNA variables often depend on each other，and how their combined status impacts individual molecules can only be assessed using longread sequencing（为什么做全长转录组，说白了就是获得完整的可变剪切信息），which sequences transcripts in single reads with no assembly required, thereby reducing alternative transcript assembly errors and enabling accurate isoform quantification.（全长的要求）。
Brain AS is especially diverse（多样性） and brain-region specific expression patterns of splicing factors（剪接因子） and other RNAbinding proteins drive brain-region-specific splicing. Examples
include diseases implicated by genes such as MAPT, Bin1,and neurexins（例子，可以不管他）。Brain-region-specific isoform expression can either originate from molecular regulation in one or multiple
cell types, or can arise purely from gene-expression or celltype abundance differences without splicing regulation.（调控与否），These distinct models are especially important during postnatal development.下面的例子我们汉语翻一下，在海马和前额叶皮层，多种细胞类型经历了分化，分化过程受发育特异性剪接的影响，不同于成熟细胞类型的剪接。However, no celltype-specific isoform investigation across brain regions exists to-date, owing to limitations in technology, throughput, and testing methods. HIPP and PFC are highly specialized regions of the telencephalon, and their circuitry is heavily implicated in movement control, cognition, learning, and memory formation. （有点专业英语的意思）。Disorders involving HIPP and PFC manifest in cognitive deficits, and understanding changes occurring at crucial developmental
timepoints of these structures is important for case–control studies.Here, we employ single-cell isoform RNA sequencing (ScISOrSeq) with increased throughput in HIPP and PFC at mouse postnatal day 7 (P7) to test and define celltype-specific contributions to brain-region-specific splicing，Furthermore, we devised a spatial isoform expression method,which provides a spatial exon expression map in addition to the existing spatial gene expression map of the Allen developing brain atlas。（多组学分析）。
ScISOrSeq这个技术作者也是第一次听说，当然，诺禾单细胞现在推出了三代全场的测序服务，我们往下看看，看看结果与方法。

（3）Results

1、Short read clustering of P7 hippocampus and prefrontal cortex tissue assigns precursors to known adult cell-types.

Our ScISOrSeq approach used barcoded single cells followed by both short and long-read analyses to reveal splice variants specific to cell types.

从技术上看，Barcode标记了之后，两种测序策略（长短两种）。短序列的测序结果用于识别细胞类型，Short-read clustering across two hippocampal replicates revealed no need for integration anchors to correct for batch effects（没有批次，牛）。然后识别细胞类型，Furthermore, we observed six vascular and immune populations including vascular endothelial cells, microglia, and macrophages。

RNA velocity analysis revealed neuronal lineages in various differentiation stages（RNA速率也做了，可惜补充材料打不开）。这里我们需要注意的是，文献中不仅仅用marker来识别细胞类型，还对数据库已经定义好的细胞类型进行Alignment，辅助识别细胞亚型，同时接住了形态学的识别。

2、A gene-wise test to determine differential isoform expression (DIE).

3、Differential isoform expression across brain regions is governed

predominantly by one specific cell type.

4、Cell types endogenous to one brain region have distinct splicing signatures.

5、Choroid plexus epithelial cells (CPEs) generate distinct isoforms predominantly through alternative TSS.

6、Single-cell basis of DIE between cell types.

When DIE is
observed between two cell types, two competing hypotheses can explain this phenomenon4. Either all cells of each cell-type behave uniformly and reflect the differences in isoform expression between the two cell types, or individual cells of one or both cell types could show variability in isoform expression. Neuronatin
(Nnat) is an important developmental gene expressing a neuronspecific isoform. In Nnat, DIE between ependymal cells and excitatory neurons is represented by the vast majority of individual cells. However, the case of DIE between excitatory neurons and granule neuroblasts is different: some granule neuroblasts
behave like excitatory neurons, while others behave like nonneurons. This may be due to different sub-populations of granule neuroblasts。

7、Clustering on long-read data recapitulates short-read cell-type assignments

我们这里主要关注几个方法：

（1）短序列的分析（cellranger和Seurat）

The 10x cellranger pipeline (version 3.0.0) was run on the raw Illumina sequencing data to obtain single-cell expression matrices. For replicate 1, the raw expression matrices obtained through cellranger were used along with the DropletUtils package (v1.6.1) to acquire ‘eligible’ barcoded single cells (FDR < = 0.001) with UMI counts that fell below cellranger’s filtering cutoff. These barcodes were incorporated into new matrices for importing into Seurat (v 3.1). For both hippocampal replicates and the first PFC replicate, cells that had unique gene counts over 5000 or less than 700, and greater than 20% mitochondrial gene expression were removed from further analysis. To adjust for the lower mean reads/cell for the second PFC
replicate, the cutoff for minimum number of genes per cell was lowered to 350. Filtering on these parameters yielded 14,433 single cells for the hippocampus across two replicates, and 10,944 single cells for the PFC. We then used Seurat’s “merge” feature to combine the replicates for each brain region. The number of UMIs, percentage of mitochondrial gene expression were regressed from each cell and then the gene expression matrix was log normalized and scaled to 10,000 reads per cell. Next, we clustered all the cells using 30 principal components (PCs) using the Louvain algorithm with a 0.6 resolution.

（2）空间短数据的分析

The 10X spaceranger pipeline was run on raw Illumina sequencing data to obtain spatial expression matrices. Seurat’s spatial analysis functions were used to obtain gene expression similarity clusters and identify barcodes corresponding to various brain regions。

（3）Integrated analysis with published data to identify cell-types.

Published RNASeq P30 mouse brain data from Allen Brain Atlas30 was used as a reference to identify cell identities of clusters based on shared gene expression patterns. Since the Allen institute data was generated using the SmartSeq2 protocol, Seurat’s integrated anchor feature77 was used to align the two datasets and transfer cell-type labels

（4）单细胞空间联合分析

P7 HIPP single-cell data was used as a reference to transfer labels onto P8 spatial transcriptomics data in the barcoded region corresponding to the hippocampus using Seurat’s integrated anchor feature using default parameters.

其他方法：有点超出本人的知识范围。但非常重要

这里大家一起努力吧，革命尚未完成，我们需要学习

本文标签： 单细胞转录全长三代文献