Skip to end of metadata
Go to start of metadata

SEARCHING THE RNA Sequencing (RNA-Seq) DATA

Three search types are available: (1) #Gene Search, to obtain gene expression data for specific genes of interest, (2) #Differential Search, to compare expression between two sets of glioblastoma anatomic structures, and (3) #Correlative Search, to find genes that have an expression pattern similar to a "seed gene" selected from the results of a Gene or a Differential search.

Clicking on the "?" button from any search type will take you to the appropriate help section.

Gene Search

With the Gene Search radio button selected (default) you can either select a gene only by typing the name of your gene of interest into the text box, or you can further restrict your selection criteria by selecting tumor features and/or donors. Once you have made your selections, click the Search button. The search will not work unless a gene has been selected. You can further limit your search by selecting the box next to "Show exact matches only".

Results will be returned in a heat map format with the genes on the y-axis (vertical) and the tumors/tumor features on the x-axis (horizontal). The heat map represents the normalized gene level RNA-Seq expression data in TPM (Transcripts per Million). By default, the columns in the heat map are sorted first by tumor, then by tumor feature. To toggle the initial sort parameter, click the toggle button in the top right hand corner of the heat map.

Clicking on a data point in the heat map will populate the metadata box above the heat map with information specific to that gene and tumor. This meta-information includes information regarding the tumor, the gene and a search box to search for genes that show similar expression patterns as your gene of interest. There are two links available from the tumor box, clicking on the "Tumor Sub-Block" link will open up a new tab with the Specimen Detail Page and clicking the "SNI" link will take you to the Swedish Neuroscience Institute page (registration is required to access the patient metadata). Gene information includes metadata on the gene selected including links to ISH data for that gene (when available) and links to other Allen Brain Atlas resources that include that gene.

If there is ISH data available for a gene, there sill be a link labeled "Yes" that will take you to the ISH Data for that gene.

You can adjust the window and the level of the color scale on the heat map by adjusting the threshold bars or re-centering the color map scale bar. You can also zoom in or out to see the heat map at different magnification levels.

You can select a set of genes for later viewing by selecting the check-boxes beside the list of gene symbols. Your choices are stored in a browser cookie in your computer and will remain in effect until you click the "Clear Selections" button or clear your web browser's cookie cache. Click the "View Selections" button to see your selections.

To restrict the columns that are displayed, select the "Filter Heatmap" button below the heatmap. Filtering your heatmap is a two step process: first, select the "..." box to restrict your features, making sure to save your selections and then toggle the filter heatmap feature by clicking the "Filter Heatmap" button.

Correlative Search

In using gene expression databases, a "search by example" feature is highly desirable as genes with similar expression patterns may be related in function. Using the "Find Correlates" search utility will accomplish this function. This search by example facility is also available in the Allen Human Brain Atlas, the BrainSpan Atlas of the Developing Brain, as well as in the Allen Mouse Brain Atlas and in the Allen Developing Mouse Brain Atlas.

Once you have selected a gene of interest (by clicking on the heat map that results from a #Gene Search or #Differential search), you can find genes with similar expression patterns. All Features and All Tumors are included in your search by default, but you can select one or more features or tumors from the drop down menus. Once you have selected the tumor feature(s) and donor(s) of interest, click "Find Correlates".

The search will return a list of genes enriched in the target domain(s) over the contrast domain(s) and ranked by it's Pearson's correlation, r.

If there is ISH data available for a gene, there sill be a link labeled "Yes" that will take you to the ISH Data for that gene.

You can adjust the window and the level of the color scale on the heat map by adjusting the threshold bars or re-centering the color map scale bar. You can also zoom in or out to see the heat map at different magnification levels.

You can select a set of genes for later viewing by selecting the check-boxes beside the list of gene symbols. Your choices are stored in a browser cookie in your computer and will remain in effect until you click the "Clear Selections" button or clear your web browser's cookie cache. Click the "View Selections" button to see your selections.

To restrict the columns that are displayed, select the "Filter Heatmap" button below the heatmap. Filtering your heatmap is a two step process: first, select the "..." box to restrict your features, making sure to save your selections and then toggle the filter heatmap feature by clicking the "Filter Heatmap" button.

You can see "anti-correlated" genes by toggling the sort order on column "r" or scrolling to the bottom of the heat map.

Differential Search

Another common usage of gene expression databases is to find genes that show enrichment of expression in one set of samples compared to another set of samples. This type of query is supported by the "Differential Search" feature. Select the "Differential Search" radio button. To find genes enriched in one "target set" of tumor features and/or donors compared to a "contrast set" of tumor features and/or donors, choose the target set using the "Target Tumor Features/Tumors" drop down menu, and the contrast set using the "Contrast Tumor Features/Tumors" drop down menu. The default setting selects all tumor features and donors, but clicking in the "All Features" or "All Tumors" text boxes will allow you to select individual tumor feature(s)/tumor(s). Then click "Search".

The search will return a list of genes enriched in the target domain(s) over the contrast domain(s) ranked by either the Fold change in expression or by the confidence (or p-Value). Clicking on either of these features will return a different list of genes. The ranking type will be indicated by a triangle next to the sort parameter.

Once selected a window will open allowing you to select the structures and the developmental stages you are interested in visualizing.

Once you have identified a single gene and want to dig deeper into the exon level expression data, click on the blue colored gene name or symbol above the heat map. This link takes you to the transcriptome visualization page.

Download Search Results

To download your search results, click the "Download this data" link at the bottom left of the heatmap. You will have the option of downloading up to 2000 genes at a time starting from any position in the heatmap. Your start position can be determined from the counter in the top left corner of the heatmap (see screenshot).

Your data will be downloaded as three separate files: 1) a Columns.csv file that lists the column headers with associated metadata for each sample, 2) a Rows.csv file that lists the row headers with associated metadata for each gene, and 3) an Expression.csv file that provides a matrix of the expression values for each data point.

The column headers, row headers and expression data will be consistent with the settings from which you downloaded the heatmap (i.e. color map).

Data Visualization

The transcriptome visualization page is divided into 3 sections.

The first section contains gene Information data including the Ensembl symbol and ID, the chromosome number, the NCBI Entrez symbol, ID and name, gene aliases, links to the USC and the Ensembl genome browsers which take you directly to gene locus data, and links to related data from other Allen Brain Atlas resources.

The second section provides a composite gene model where exons are represented by boxes and introns are represented by the green lines between exons. When you hover your mouse over an exon in either the RNA-Seq or exon array data, the exon is highlighted in this model.

The third section contains two tabs:
A tab labeled RNA-Seq with a heat map of the exon level expression data in RPKM (Reads Per Kilobase of exon model per Million mapped reads).
A second tab labeled Exon Array with a heat map of the exon probeset level expression data in log2 transformed normalized expression intensity values.

This data can be downloaded by clicking on the "Download this data" link at the bottom left of the heatmap.

Transcriptome Heat Maps

The heat map is a visualization of the exon expression values for the returned gene of interest. The heat map data is presented as a matrix with brain structure (by developmental stage) on the x-axis and gene exons on the y-axis. Brain structures are organized in ontological order. Clicking on the toggle button in the right hand corner will toggle the initial sorting parameter from structure to developmental stage.

Each row of the heat map in the RNA-Seq data represents an exon. Exons in the RNA-Seq data are labeled by the start position on the chromosome and the length of the exon. Each row of the heat map in the Exon Array data represents an exon probeset. Approximately 4 probes are selected for each putative exon region and called a probeset. Each row in the Exon Array data is labeled with the chromosome start position and the length of the probeset. Hovering the mouse over the exon in either heat map will highlight the corresponding exon in the composite gene model over the array data.

Each column of the heat map represents a tissue sample. The colors of the heat map are expression values, transformed to a log 2 scale. The color scale ranges from dark blue representing low expression and passes through cyan, yellow, orange and finally to dark red representing high expression. Hovering the mouse over a cell of the heat map will bring up donor and structure information in the area above the heat map.

  • No labels