Getting Started with the Transcriptomics Explorer

[tylermo]

Overview
The Transcriptomics Explorer is a browser-based tool for researchers to visualize and analyze scRNA sequencing data and related cell types for both mouse and human. A previous version of this tool - called the RNA-seq Data Navigator - is still available for both mouse and human.

Currently researchers can:

• Explore a gene expression heatmap across cell types in the dataset, and search for genes of interest to inspect
• Explore a tSNE or UMAP visualization of all cells in the dataset, colored by cell types or gene expression
• Visualize a dataset’s sampling strategy to see how cells and nuclei were sampled across brain areas, cortical layer, and other dimensions
• Find a cell type of interest in one visualization and see its characteristics in a different visualization
• Download data files, including cell-level gene expression and taxonomic metadata and structure

Toggle Datasets
Toggle between datasets by clicking on the menu in the black bar near the top left corner of the screen. This post lists the available datasets.

Toggle Visualizations
Toggle between the three visualizations by clicking on the “Visualizations” menu in the white bar near the top left corner of the screen.

To get started with each visualization, see these posts:

• Heatmap
• Scatterplot
• Sampling Strategy

Download Data
Several files are available for download, including cell-level gene expression, cell type taxonomy information, and cell metadata and annotations. Click “Download Data” in the bottom left part of the screen and choose the files of interest from the Download table.

Currently, the entire dataset must be downloaded as a whole.

Cell Type Nomenclature & Taxonomy
The Cell Types nomenclature is described here: Cell Type Nomenclature - brain-map.org

[cjohnson23]

Hello! This is all very helpful. I am a high school student learning about neuroscience. Currently, I am using the RNA patch seq data to look for correlation data between gene expression and specific metrics of neuronal morphology. I did notice that some genes in the transcriptomics file had an expression level of 0.0cpm in some cells. As a result, the correlation tests I have conducted have been skewed. I was wondering what exactly these 0.0 values mean. Does this mean there is no expression of that gene in the cell? Or does this indicate a null value? Could I remove those 0 values from the dataset when conducting correlation tests, or would that have bad implications? Please let me know. Thanks.

[jeremyinseattle]

Welcome, and we’re glad to hear you find this resource useful! A CPM value of 0 is very meaningful–as you suggest, it means that there is no expression of that gene in that cell. I would strongly advise against removing those 0 values from the dataset by setting them as “NA” values. However, you should remove the genes where EVERY cpm value = 0. Retaining these genes will cause problems in many analyses.

Another option, that we typically use, is to restrict your analysis only to genes that are expressed in some cell types but not others. There are many ways to do this, but the least complicated is to sort the genes by variation (or standard deviation) across cells, and then consider only the top 5,000 or so most variable genes.