Deciphering Brain Complexity Using Single-cell Sequencing

Abstract

The human brain contains billions of highly differentiated and interconnected cells that form intricate neural networks and collectively control the physical activities and high-level cognitive functions, such as memory, decision-making, and social behavior. Big data is required to decipher the complexity of cell types, as well as connectivity and functions of the brain. The newly developed single-cell sequencing technology, which provides a comprehensive landscape of brain cell type diversity by profiling the transcriptome, genome, and/or epigenome of individual cells, has contributed substantially to revealing the complexity and dynamics of the brain and providing new insights into brain development and brain-related disorders. In this review, we first introduce the progresses in both experimental and computational methods of single-cell sequencing technology. Applications of single-cell sequencing-based technologies in brain research, including cell type classification, brain development, and brain disease mechanisms, are then elucidated by representative studies. Lastly, we provided our perspectives into the challenges and future developments in the field of single-cell sequencing. In summary, this mini review aims to provide an overview of how big data generated from single-cell sequencing have empowered the advancements in neuroscience and shed light on the complex problems in understanding brain functions and diseases.

Keywords: Brain development; Brain diseases; Cell type; Neuroscience; Single-cell RNA-seq.

Figure 1

A typical workflow of scRNA-seq data analysis The workflow consists of six steps. Step 1: preprocessing, in which the raw sequencing data are cleaned, demultiplexed, mapped to the reference genome, and quantified. The output of this step is a gene expression matrix. Step 2: normalization, in which the raw expression data are normalized to denoise and remove batch effects. Step 3: dimensionality reduction, in which the high dimension data are projected to a small number of dimensions to capture the main signal. Step 4: clustering, in which the cells are assigned to clusters, which may represent different cell types or states. Step 5: differential gene expression, in which comparisons are performed between cells of different clusters or from different groups. The output of this step is a list of differentially-expressed genes. Step 6: gene expression dynamics, in which a developmental trajectory connecting different cell clusters is inferred from the expression patterns. Exemplary tools are listed for each step. UMI, unique molecular identifier.

Figure 2

The exponential increase of the number of cells sequenced in published scRNA-seq studies of the brain. The number of published scRNA-seq studies of the brain (as of August 30, 2019) we manually found is shown in the top panel. The number of sequenced cells in each study is shown in the bottom panel. Each circle stands for one study, and the exponential trend of the number of sequenced cells was fitted by robust linear regression, with 95% confidential interval shown in gray.