Loading...

Welcome to CoDEx Viewer!

(Cortical Development Expression Viewer)
Getting Started: See help items below or choose an option from drop down menu

Welcome to CoDex (Cortical Development Expression) viewer! This resource comprises a single-cell expression dataset of the developing human neocortex. Here you can explore the dataset using heatmaps, tSNE’s, distribution graphs and statistical summary tables. In addition, you can explore which genes are differentially expressed for a given cell type during cortical development.

Schematic of the developing human neocortex. VZ: ventricular zone; iSVZ: inner subventricular zone; oSVZ: outer subventricular zone; IZ: intermediate zone; SP: subplate; CPi: inner cortical plate; CPo: outer cortical plate; RG: radial glia; IP: intermediate progenitor; MN: newborn migrating excitatory neuron; EN: excitatory neuron; IN: interneuron; O: oligodendrocyte precursor; E: endothelial cell; P: pericyte; M: microglia. Scatter plot visualization of cells after t-stochastic neighbor embedding (tSNE), colored by cell type annotation.

This dataset was generated by Geschwind Lab. This webtool was built using R Shiny by Andrew Elkins. To view our associated publication:

A Single-Cell Transcriptomic Atlas of Human Neocortical Development during Mid-gestation

Damon Polioudakis†, Luis de la Torre-Ubieta†, Justin Langerman, Andrew G. Elkins, Xu Shi, Jason L. Stein, Celine K. Vuong, Susanne Nichterwitz, Melinda Gevorgian, Carli K. Opland, Daning Lu, William Connell, Elizabeth K. Ruzzo, Jennifer K. Lowe, Tarik Hadzic, Flora I. Hinz, Shan Sabri, William E. Lowry, Mark B. Gerstein, Kathrin Plath, Daniel H. Geschwind

Abstract

We performed RNA sequencing on 40,000 cells to create a high-resolution single-cell gene expression atlas of developing human cortex, providing the first single-cell characterization of previously uncharacterized cell types, including human sub-plate neurons, comparisons with bulk tissue, and systematic analyses of technical factors. These data permit deconvolution of regulatory networks connecting regulatory elements and transcriptional drivers to single-cell gene expression programs, significantly extending our understanding of human neurogenesis, cortical evolution, and the cellular basis of neuropsychiatric disease. We tie cell-cycle progression with early cell fate decisions during neurogenesis, demonstrating that differentiation occurs on a transcriptomic continuum; rather than only expressing a few transcription factors that drive cell fates, differentiating cells express broad, mixed cell-type transcriptomes before telophase.By mapping neuropsychiatric disease genes to cell types, we implicate dysregulation of specific cell types in ASD, ID, and epilepsy. We developed CoDEx, an online portal to facilitate data access and browsing.

The heatmap tool allows you to query genes expressed during neocortical development and generates a heatmap with mean normalized expression for the given cell types present. The query accepts HGNC gene names with the ability to auto-fill the gene name by pressing enter. After entering a gene name, press the Enter/Return key to begin entering a new gene. The search bar allows you to search up to 30 genes, use the Tab key to exit the search box. Another option is to copy and paste a comma separated list of genes. These genes must not have quotes around them.
With the tSNE tool you can visualize the expression of individual cells plotted on our clustered tSNE graph. For more information about tSNE see this link. Here are the basics: each point on the graph represents a cell and the relative closeness of cells to each other tells us about the underlying similarity between the cells. A key illustrating the cell types of each cell is provided under the search bar. When querying a gene, the normalized expression level is given via white to red gradient.
Using the distribution tool, you can find the normalized expression levels of specific genes across all of the cell types plotted as either a boxplot or histogram.
For multiple genes you can query for summary statistics for each gene. Statistics are given for each cell type and across all cells in a searchable table.
Finally, you can also look to see which genes are enriched in each cell type. When clicking on a cell type, a heatmap with genes sorted by expression level is generated. In addition, genes that are enriched in this cell type are given in a searchable table to the right of the heatmap.
Normalization was performed using Seurat-v2.0.1. Briefly, raw counts are read depth normalized by dividing by the total number of UMIs per cell, then multiplying by 10,000, adding a value of 1, and log transforming (ln (transcripts-per-10,000 + 1)) using the Seurat function ‘CreateSeuratObject’. Raw UMI counts data were assessed for the effects from biological covariates (anatomical region, donor, age, sex), and technical covariates (library batch, sequencing batch, number of UMI, number of genes detected, CDS length, GC content) (fig. S1). The effects of number of UMI (sequencing depth), donor, and library preparation batch were removed using a linear model from the read depth normalized expression values (custom R scripts, lm(expression ~ number_of_UMI + donor + lab_batch), and Seurat function "ScaleData."
Help

Download help doc here

Feedback

If you have feedback or want to report a bug please email aelkins[at]mednet.ucla.edu

Download Graph
Download Graph

Data Download

The download contains the raw counts (UMI) gene expression matrix and metadata:

Download data

The count matrix (raw_count_mat) is stored as a sparse matrix in an rdata file. To convert to matrix use the "Matrix" package using R.

 library(Matrix)
                                        
 # load in sparse matrix
 load("../sc_dev_cortex_geschwind/raw_counts_mat.rdata")
 raw.counts.mat <- as.matrix(raw_counts_mat)

The raw Drop-seq data was processed using the Drop-seq tools v1.12 pipeline from the McCarroll Laboratory. Reads were aligned to the Ensembl release 87 Homo sapiens genome. We calculated unique molecular identifier (UMI) counts for each gene of each cell by collapsing UMI reads using Drop-seq tools.

To select Drop-seq cells for downstream analysis:

1) Cells were selected for downstream analysis using the cell barcodes associated with the most UMIs. We estimated the number of cells captured as 5% of the input beads and retained this many cell barcodes for downstream analysis.

2) For samples with mouse cells spiked in, mouse cells were removed by filtering all cells with > 250 UMIs mapping to the mouse genome.

3) Removed cells with <200 unique genes detected (gene detection: >=1 count).

4) Removed cells with >3 standard deviations above the mean number of genes detected (3152).

5) Removed cells with >5% of their counts mapping to MT genes.

6) Removed genes detected in <3 cells.