The Allen Mouse Brain Atlas provides genome-wide in situ hybridization (ISH) image data for approximately 20,000 genes in adult mice. Each data set is processed through an informatics analysis pipeline to obtain spatially mapped quantified expression information.

From the API, you can:

• Download images
• Download quantified expression values by structure
• Download quantified expression values as 3-D grids
• Query the differential and correlative search services
• Query the image synchronization service
• Download atlas images, drawings and structure ontology

This document provides a brief overview of the data, database organization and example queries. API database object names are in camel case. See the main API documentation for more information on data models and query syntax.

‍

Experimental Overview And Metadata

Experimental data from this atlas is associated with the “Mouse Brain” Product.

Multiple genes were assayed using each Specimen. Typically, the sectioning scheme divided each brain into eight interleaving SectionDataSets with 200 µm sampling density (= 8 x 25 µm thickness).

Each gene was assayed with at least one sagittal SectionDataSet. A subset of genes also has a coronal SectionDataSet and/or replicate experiments. A sagittal SectionDataSet spans the left hemisphere starting with the most lateral section located at where the hippocampus starts to appear to just past the midline yielding ~20 SectionImages at 200 µm sampling density. A coronal SectionDataSet spans both hemispheres starting with the most posterior section showing the cerebellum and hindbrain to the most anterior section showing the olfactory bulb, yielding ~60 SectionImages at 200 µm sampling density. The left side of a coronal SectionImage corresponds to the left hemisphere.

A manual QC protocol defines the criteria for failing experiments due to production issues, discarding damaged SectionImages, verifying and adjusting the tissue bounding boxes, as well as for identifying “dark” artifacts such as bubbles and tears.

From the API, detailed information about Genes, Probes, SectionDataSets and SectionImages can be obtained using RMA queries.

Examples

• All genes in the “Mouse Brain” Product
• All experiments associated with gene prodynorphin (Pdyn)

From the above query, gene Pdyn has one sagittal SectionDataSet (id=69782969) and one coronal SectionDataSet (id=71717084). In the web application, images from the experiment are visualized in an experiment detail page. All displayed information, images and structural expression values are also available through the API.

See the image download page to learn how to download images at different resolutions and regions of interest.

Informatics Data Processing

The informatics data processing pipeline produces results that enable the navigation, analysis and visualization. The pipeline consists of the following components:

• an annotated 3-D reference space,
• an alignment module,
• an expression detection module,
• an expression gridding module, and
• a structure unionizer module.

The output of the pipeline is quantified expression values at a grid voxel level and at a structure level according to the integrated reference atlas ontology. The grid level data are used downstream to provide a differential and correlative gene search service and to support visualization of spatial relationships. See the informatics processing whitepaper for more details.

‍

3-D Reference Models

The backbone of the automated pipeline is an annotated 3-D reference space based on the same Specimen used for the coronal plates of the integrated reference atlas. A brain volume was reconstructed from the SectionImages using a combination of high frequency section-to-section histology registration with low-frequency histology to (ex-cranio) MRI registration. This first-stage reconstructed volume was then aligned with a sagittally sectioned Specimen. Once a straight mid-sagittal plane was achieved, a synthetic symmetric space was created by reflecting one hemisphere to the other side of the volume.

Over 800 Structures were extracted from the 2-D coronal reference atlas plates and interpolated to create symmetric 3-D annotations. Structures in the reference atlas are arranged in a hierarchical organization. Each structure has one parent and denotes a “part-of” relationship. Structures are assigned a color to visually emphasize their hierarchical positions in the brain.

In the May 2015 data release, we introduced a next generation common coordinate framework (CCF v3) based on a population average to support the integration of new mouse brain datasets in the Allen Brain Atlas Data Portal. See the Allen Mouse Common Coordinate Framework whitepaper for detailed construction information.

The Nissl volume and 3-D annotation from the Allen Reference Atlas were deformably registered to the new common coordinate framework to support potential cross-modality analysis of gene expression with new data modalities as they become available in the Data Portal.

‍
See the atlas drawings and ontologies page for more information.

All coronal data is registered to ReferenceSpace id = 9. All sagittal data is registered to ReferenceSpace id = 10.

ReferenceSpace id = 9 is in PIR orientation (+x = posterior, +y = inferior, +z = right). ReferenceSpace id = 10 is identical to ReferenceSpace id = 9. The reason for the two spaces is to allow left hemisphere sagittal data to correspond to the right hemisphere coronal reference atlas. This is implemented as a z-axis flip transform between the two reference spaces for the purposes of image synchronization (see below).

NOTE: 3-D annotation volumes were updated in the May 2015 release to reflect the introduction of the next generation of the Allen Mouse Common Coordinate Framework (CCFv3). Annotation volumes from the October 2014 release (mapped to CCFv2) can be access through our data download server (see instructions).

Three volumetric data files are available for download:

• atlasVolume : uchar (8bit) grayscale Nissl volume of the reconstructed brain at 25 µm resolution.
• annotation : uint (32bit) structural annotation volume at 25 µm resolution. The value represents the ID of the finest level structure annotated for the voxel. Note: the 3-D mask for any structure is composed of all voxels annotated for that structure and all of its descendents in the structure hierarchy.
• gridAnnotation : uint (32bit) structural annotation volume at grid (200 µm) resolution for gene expression analysis.

All volumetric data is stored in an uncompressed format with a simple text header file in MetaImage format. The raw numerical data is stored as a 1-D array as shown in the figure below.

‍

Example Matlab code snippet to read in the 25µm atlas and annotation volume:

‍

% Download and unzip the atlasVolume and annotation zip files

% 25 micron volume size

size = [528 320 456];

% VOL = 3-D matrix of atlas Nissl volume

fid = fopen('atlasVolume/atlasVolume.raw', 'r', 'l' );

VOL = fread( fid, prod(size), 'uint8' );

fclose( fid );

VOL = reshape(VOL,size);

% ANO = 3-D matrix of annotation labels

fid = fopen('annotation.raw', 'r', 'l' );

ANO = fread( fid, prod(size), 'uint32' );

fclose( fid );

ANO = reshape(ANO,size);

% Display one coronal section

figure;imagesc(squeeze(VOL(264,:,:)));colormap(gray);

figure;imagesc(squeeze(ANO(264,:,:)));colormap(lines);

% Display one sagittal section

figure;imagesc(squeeze(ANO(:,:,220)));colormap(lines);

figure;imagesc(squeeze(VOL(:,:,220)));colormap(gray);

‍

Example Matlab code snippet to read in the 200µm atlas and annotation volume:

‍

% Download and unzip the gridAnnotation zip files

% 200 micron volume size

sizeGrid = [67 41 58];

% ANOGD = 3-D matrix of grid-level annotation labels

fid = fopen( 'gridAnnotation.raw', 'r', 'l' );

ANOGD = fread( fid, prod(sizeGrid), 'uint32' );

fclose( fid );

ANOGD = reshape(ANOGD,sizeGrid);

% Display one coronal and one sagittal section

figure;imagesc(squeeze(ANOGD(34,:,:)));colormap(lines);

figure;imagesc(squeeze(ANOGD(:,:,28)));colormap(lines);

‍

Image Alignment

The aim of image alignment is to establish a mapping from each SectionImage to the 3-D reference space. The module reconstructs a 3-D Specimen volume from its constituent SectionImages and registers the volume to the 3-D reference model by maximizing image correlation.

Once registration is achieved, information from the 3-D reference model can be transferred to the reconstructed Specimen and vice versa. The resulting transform information is stored in the database. Each SectionImage has an Alignment2d object that represents the 2-D affine transform between an image pixel position and a location in the Specimen volume. Each SectionDataSet has an Alignment3d object that represents the 3-D affine transform between a location in the Specimen volume and a point in the 3-D reference model. Spatial correspondence between any two SectionDataSets from different Specimens can be established by composing these transforms.

For convenience, a set of “Image Sync” API methods is available to find corresponding position between SectionDataSets, the 3-D reference model and structures. Note that all locations on SectionImages are reported in pixel coordinates and all locations in 3-D ReferenceSpaces are reported in microns. These methods are used by the Web application to provide the image synchronization feature in the multiple image viewer (see Figure).

Examples:

• Fetch alignment transforms parameters for the sagittal Pdyn SectionDataSet
• Sync a location between the sagittal and coronal Pdyn SectionDataSets

‍

‍

Expression Detection

For every ISH SectionImage, a grayscale mask is generated that identifies pixels corresponding to gene expression. The detection algorithm is based on adaptive thresholding and mathematical morphology.

The expression mask image is the same size and pixel resolution as the primary ISH image and can be downloaded through the image download service.

‍

‍

Expression Gridding

For each SectionDataSet, the Gridding module creates a low resolution 3-D summary of the gene expression and projects the data to the common coordinate space of the 3-D reference model. Casting all data into a canonical space allows for easy cross-comparison of gene expression data from every Product. The expression data grids can also be viewed directly as 3-D volumes or used for analysis (i.e. differential and correlative searches).

Each image in a SectionDataSet is divided into a 200 x 200 µm grid. Pixel-based gene expression statistics are computed using information from the primary ISH and the expression mask:

• expression density = sum of expressing pixels / sum of all pixels in division
• expression intensity = sum of expressing pixel intensity / sum of expressing pixels
• expression energy = expression intensity * expression density

Each per-image 2-D expression grid is smoothed and rotated to form a 3-D grid. Z-direction smoothing is applied to the 3-D grid which is then transformed into the standard reference space.

Grid data can be downloaded for each SectionDataSet using the 3-D Expression Grid Data Service. The service returns a zip file containing the volumetric data for expression density, intensity and/or energy in an uncompressed format with a simple text header file in MetaImage format. Structural annotation for each grid voxel can be obtained via the ReferenceSpace gridAnnotation volume file.

Note: while the reference space spans both hemispheres, sagittal SectionDataSets only span the left hemisphere. Voxels with no data are assigned a value of “-1”.

Examples:

• Download expression energy grid file for the coronal Pdyn SectionDataSet
• Download expression density and intensity grid files for the same SectionDataSet

NOTE: Grid data were updated in the May 2015 release to reflect the introduction of the next generation of the Allen Mouse Common Coordinate Framework (CCFv3). Grid data from the October 2014 release (mapped to CCFv2) can be access through our data download server (see instructions).

The expression data grid can be viewed in the Brain Explorer® 2 desktop program. Each grid voxel is rendered as a colorized sphere where the diameter represents expression energy and the color encoding expression intensity. In addition, a preview of the expression data grid is shown on the Web application as a series of maximum density projection images.

Example Matlab code snippet to read in the 200 µm energy grid volume:

‍

% Download and unzip the energy grid file for Pdyn SectionDataSet

% 200 micron volume size

sizeGrid = [67 41 58];

% ENERGY = 3-D matrix of expression energy grid volume

fid = fopen('Pdyn_P56_coronal_71717084/energy.raw', 'r', 'l' );

ENERGY = fread( fid, prod(sizeGrid), 'float' );

fclose( fid );

ENERGY = reshape(ENERGY,sizeGrid);

% Display one coronal and one sagittal section

figure;imagesc(squeeze(ENERGY(34,:,:)));colormap(gray);

figure;imagesc(squeeze(ENERGY(:,:,28)));colormap(gray);

‍

Structure Unionization

Expression statistics can be computed for each structure delineated in the reference atlas by combining/unionizing grid voxels with the same 3-D structural label. While the reference atlas is typically annotated at the lowest level of the tree, statistics at upper level structures can be obtained by combining measurements of the hierarchical children. This process produces expression density, intensity and energy measurements for each experiment and structures of interest.

Expression statistics are encapsulated as a StructureUnionize object associated with one Structure and one SectionDataSet and can be downloaded via RMA.

Example:

• Download structure expression values for the sagittal Pdyn SectionDataSet

StructureUnionize data is used in the web application to display expression summary bar graphs for a set of coarse structures.

‍

Expression Grid Search Service

A expression grid search service has been implemented to allow users to instantly search over the ~25,000 SectionDataSets to find genes with specific expression patterns:

• The Differential Search function allows users to find genes which have higher expression in one structure (or set of structures) compared to another structure (or set of structures).
• The Correlation Search function enables the user to find genes that have a similar spatial expression profile to a seed gene when compared over a user-specified domain.

The expression grid search service is available through both the Web application and API.

To perform a Differential Search, a user specifies a set of target structures and a set of contrast structures. In the service, the set of voxels belonging to any of the target structures forms the target voxel set, and voxels belonging to any of the contrast structures form the contrast voxel set. For each SectionDataSet a fold change is computed as the ratio of average expression energy in the target voxel set over the average expression energy in the contrast voxel set. The return list is sorted in descending order by fold-change.

Example: Differential search for genes with higher expression in the thalamus than the isocortex

• Pipe1: Set up the contrast structure list by finding the structure isocortex within the Mouse Brain ontology
• Pipe2: Set up the target structure list by finding the structure thalamus within the Mouse Brain ontology
• Connect the two pipes to service::mouse_differential to perform the differential search
• Visualize the same search result in the web application

See the connected service page for definitions of service::mouse_differential parameters.

To perform a Correlation search, a user selects a seed SectionDataSet and a domain over which the similarity comparison is to be made. All voxels belonging to any of the domain structures form the domain voxel set. Pearson’s correlation coefficient is computed between the domain voxel set from the seed SectionDataSet and every other SectionDataSet in the Product. The return list is sorted by descending correlation coefficient.

Example: Correlation search for genes with similar expression to the sagittal Pdyn SectionDataSet

• Pipe: Set up the seed SectionDataSet by finding the sagittial SectionDataSet for gene Pdyn
• Connect the pipe to service::mouse_correlation to perform the correlation search
• Visualize the same search result in the Web application

See the connected service page for definitions of service::mouse_correlation parameters.

‍

In order to perform these computations quickly over the entire data set, a subset of voxels are loaded in memory. The full expression grid is 67x41x58=159,326 voxels spanning both hemispheres and includes background voxels. To load all voxels for all image series into memory would require 14GB of RAM. To reduce memory requirements and increase the efficiency of calculations, voxels spanning over 80% of all experiments were identified. Only these ~26,000 voxels were then used in the ‘‘full’’ search service requiring 4 GB of RAM and partially spanning one hemisphere.

To take advantage the data on both hemispheres in coronal data, a second ‘‘coronal only’’ search service is also available as an option. The coronal service spans both hemispheres covering 58,387 voxels and searches over the ~4,000 coronal image series.

It should be noted that this on-the-fly search service is derived from a fully automated processing pipeline. False positive and false negative results can occur due to artifacts on the tissue section or slide and/or algorithmic inaccuracies. Users should confirm results with visual inspection of the ISH images.

‍

‍

Viewing Gene Expression Data

The Allen Mouse Brain Atlas offers the following ways for you to find gene expression data:

• Use Gene Search to locate genes’ data by name, symbol, NCBI accession number, or Entrez Gene ID.
• You can also use Gene Search with Boolean operators on specific field names using Boolean Syntax Query.
• Search for genes with enhanced expression by anatomic region using Differential Search.
• Search for manually curated enhanced gene expression data from 75 small structures using the Fine Structure Annotation.
• Differential Search across both the mouse and human gene expression datasets using the Human Differential Search.
• Search Related Studies for gene expression in our Sleep Study or our Mouse Strains Study.

‍

Gene Search

To search for a specific gene by name, symbol, NCBI accession number, or Entrez gene ID, type your query into the text box, making sure the “Gene Search” radio button is selected. As you type, genes that match your search string will be suggested to you. Select an item from the list and click the “Search” button. The results are displayed in a list that shows all genes with expression data that match your query.

Tips for searching:

• If you are unsure of spelling, begin typing the gene identifier and suggestions will be offered.
• Press the ‘Enter’ key instead of clicking the “Search” button.
• Include an asterisk (*) as a wildcard character before or after your search text to return genes that start with or end with that text. For example, enter pig* to return genes that start with pig or *pig to return genes that end with pig.
• If you do not include an asterisk (*) wildcard character, wildcards are implied to surround your search text. So if you search using the text pig, for example, you’ll get results that include pig anywhere within the gene name or symbol.
• Select the “Show exact matches only” check box to remove the implied wildcards. This is useful if you want to return only those genes that exactly match your search text.
• Place the search text within quotes to include spaces or special characters.
• Click the “Bulk Search” radio button to search for many genes at once. Doing so expands the text box, making it easier to copy and paste a list of genes from a text file or spreadsheet. Comma and semi-colon are not valid separators for genes; use tab or | instead. The bulk query does not implicitly include wildcard characters around your search text (you must specify them).

‍

Boolean Syntax Query

The following special operators can be used to build queries:

• AND, OR and NOT must be upper case.
• The AND function is implicit, so spaces between words that are not separated by an operator will be treated like AND.
• OR has higher operator precedence than AND.
• Parentheses can be used to group criteria, but nested parentheses are not supported at this time.

‍

Differential Search

The Differential Search option is ideal to search for gene expression data when the brain structure or area is known but little is known about genes that may be expressed in that area. To find genes with enhanced expression in one or more brain regions in contrast to one or more different regions, first select the “Differential Search” radio button. Enter the target structure(s) in the top search box labeled “Target Structure(s)” with a comma separating each structure. You can then enter “Grey” (for grey matter) or some smaller region as your contrast region in the “Contrast Structure(s)” search box, then click on the Search button.

The results are displayed as a list of genes with enhanced expression in the regions that match your query.

Differential search is an on-the-fly search service. Calculations over all ~25 000 experiments are computed as a fold change using expression energy data from one hemisphere. To restrict the search to coronal experiments (~4 000) which used data from both hemispheres, select the “Coronal data only” check box.

Since this search service is derived from a fully automated processing pipeline, false positive and false negative results can occur due to artifacts on the tissue section or slide and/or algorithmic inaccuracies. Users should confirm results with visual inspection of the ISH images. The “Expression threshold ” slider may be useful to reduce false positive results due to dark spot artifacts. This threshold restricts the list of returned results by excluding any experiment where the expression energy of the target structures is less than threshold.

‍

Fine Structure Annotation

A differential search with manually curated results was generated for 75 small structures in the brain listing the top 50 genes with specific and enhanced gene expression. To use the Fine Structure search, select the “Fine Structure Annotation” radio button, then select a brain region from the drop down menu to return this list.

‍

Human Differential Search

The “Human Differential Search” function enables side by side viewing and comparison of differential search across the mouse ISH and human microarray datasets. Select the “Human Differential Search” radio button. To find genes with higher expression in one structure compared to another structure, select a target brain structure in the top drop-down box and a contrast region in the bottom drop-down box and click on the Search button.

‍

Search results

The return list is generated by first executing a differential search query in the mouse ISH dataset. Experiments for each gene are grouped together and ordered by maximum fold-change. For each gene in the mouse ISH return list, associated experiments in the human microarray dataset are identified; fold-change values are computed and displayed. Note that NCBI HomoloGene data is used to link genes across organisms. Click on the dataset column heading to toggle the primary dataset selection.

For each experiment, a gestalt visualization of the brain wide expression pattern is also provided. The type of visualization depends on the dataset:

For mouse brain data, expression is visualized as rotating 3D thumbnails. Each view is generated by maximum density projection where denser expression area appears more solid and obscures areas of lower density. Each cube is color-coded by average expression intensity ranging from blue (low intensity) through green to red (high intensity). Move slider bar in the header to rotate thumbnail.

For human brain data, microarray data is visualized as a heatmap where each column represents a tissue sample in ontological order. Heatmap color represents the z-score over a probe ranging from green (z-score of -3 and below) through black to red (z-score of +3 and above). Hover over any position in the heatmap to obtain the associated tissue sample structural annotation.

‍

Correlative Search

In using gene expression databases, a “search by example” feature is also highly desirable as genes with similar expression patterns may be related in function. The Correlative Search utility will accomplish this function. This search by example facility is also available in the Allen Human Brain Atlas, Allen Developing Mouse Brain Atlas and the BrainSpan atlas of the developing human brain.

Once you have identified a gene of interest, to find other genes with spatial expression profiles similar to your gene of interest, first select your experiment by clicking on the gene name in the search results list. You will see that the experiment loads in the panel to the right hand side of the results list. Select the brain structure(s) in which you would like to see a similar expression pattern from the drop-down menu. You can search all the data or just the coronal data, by clicking the “Search Coronal Data Only” box, and click “Search”. This action will return experiments with a similar expression profile to brain region(s) in which you are interested.

‍

Viewing Search Results

Searches will return a list of experiments based on the input search criteria. Each row includes the following information.

Select the experiments to view in greater detail by clicking on the checkboxes next to your experiments of interest. Experiments will be saved in your cache until you select the “Clear Selections” button at the bottom of the page. After selecting one or more experiments, click the “View Selections” button.

You can look for genes with a similar expression profile to your gene of interest by using the Correlative Search option: once a gene is selected, choose a structure from the drop-down menu and click “Search”. You can limit your results to only the data collected in the coronal plane by selecting the “Coronal data only” radio button.

‍

Experiment Image Viewer

The experiment image viewer displays images for each selected experiment in a Zoom and Pan viewer. This view makes it easy to compare experiments with each other and with the Reference Atlas.

Multiple image series can be opened on the same page to enable side-by-side comparisons. Arrange the experiments by dragging a viewer by the title bar and dropping in a new location. Add a reference atlas by selecting one in the “Atlases” drop-down menu in the upper-right hand corner of the window.

If you are viewing more than one experiment, open the configuration options to change the number of columns displayed in the window. The configuration options are accessible by clicking on the button with a “gear” icon to the right of the “Atlases” menu.

The gene symbol is displayed in the title bar along with the Allen Institute image series ID. Additional details are displayed across the top of the viewing area, including the image index, tissue index, and tissue location.

Thumbnails for the entire image series are displayed across the bottom of the viewer in section order. Click a thumbnail to select it for viewing, or use the keyboard to navigate through the set. The current selection is outlined in black.

‍

Experimental Detail

When the experiment is selected by clicking on the link, a summary of the experimental detail is returned. The data includes 1) an interactive 3-dimensional representation of gene expression, 2) a histogram of the expression energy in 12 different structures, 3) probe and gene metadata and 4) an interactive image viewer that displays the images in the experiment.

Hovering your mouse over the histogram will sync the image in the single image viewer with the section corresponding to that structure and expression.

‍

Using the Zoom and Pan (ZAP) Image Viewer

The Zoom and Pan (ZAP) Image Viewer is a powerful tool to navigate and view the images in an experiment. The main part of the viewer is an interactive window where an image can be repositioned by dragging with a mouse. Use the scroll wheel or the keyboard to zoom in or out.

Select other images in the experiment by clicking on a thumbnail image below the main viewer.

‍

Scale Bar

Drag the scale bar with your mouse to the desired location. Click on the text with your mouse to toggle between horizontal and vertical.

‍

Using the ZAP Viewer Toolbar

Use the toolbar to take actions on the image that currently has focus. Toolbar controls include:

Control	Function
	Move the slider bar to the left to intensify the contrast or color
	Select ISH, Nissl or Expression image type
	Display all thumbnail images in a single ‘contact sheet’
	Open the selected image in the image viewer tool
	Synchronize all other Zap viewers on the page that support synchronization to the currently selected image
	Open a new browser window to review metadata for the image series
	Close the Zap viewer

‍

Keyboard Commands

Use the keyboard to navigate through the image series and synchronize the viewers on the page. Keyboard commands include:

You can also use the arrow keys to pan the current image.

‍

Expression Energy

The expression mask image display highlights those cells that have the highest probability of gene expression using a heat map color scale (from low/blue to high/red).

The Expression Energy was calculated as follows: Within a given area A (voxel or structure), expression energy = (sum of intensity of expressing pixels in A) / (sum of all pixels in A)

‍

Using the High Resolution Image Viewer

The image viewer enables you to view a high resolution image in its own re-sizable window. It is possible to open multiple images by opening multiple windows.

Ways to interact with an image include:

• Use the on screen navigation buttons to zoom, pan and move forward or backward through the image series. You can also use the Keyboard Commands used in the ZAP Image viewer.
• Drag the scale bar with your mouse to the desired location. Click the text with your mouse to toggle between horizontal and vertical.
• Use the Configuration Menu to download images and adjust image parameters.
• Use the Interactive Atlas Viewer to view the reference atlas along side the ISH images

What Is AGEA?

The Allen Institute has built a data-driven three-dimensional atlas of the adult C57Bl/6J mouse brain based on the ISH gene expression images of the Allen Mouse Brain Atlas. Only the coronal gene expression data were used to construct this application. The AGEA feature (which stands for Anatomic Gene Expression Atlas) characterizes the multi-scale spatial relationship in the mouse brain as derived from coronal gene expression data without a prior knowledge of classical anatomy. Using the AGEA feature of the Allen Mouse Brain Atlas, you can:

• View and navigate 3-D spatial relationship maps (Correlation mode) and search for genes within the coronal dataset with local regionality (Find Genes) and,
• Explore a transcriptome based spatial organization of the brain (Clusters mode).

To launch this feature, click on the AGEA tab from the Allen Mouse Brain Atlas and you will see the following image.

The upper panel of images are orthogonal views of a 3-D Nissl reference atlas volume. The reference space was divided into 200 µm volumes, or voxels. The cross hairs in the top panel of images select for a seed voxel from which to compare to gene expression profiles to all other voxels in the brain. For more detailed information on how AGEA was constructed please see the AGEA user guide or Ng L, et al.(2009) An anatomic gene expression atlas of the adult mouse brain. Nature Neuroscience 12(3): 356-362.

‍

The AGEA Viewer

This section describes the controls for the Correlation and Cluster modes of AGEA.

A. Seed Selector panel: Use this area to select the seed voxel as marked by the red crosshairs. Either click or click and drag to navigate to a different correlation map.
B. Map panel: This area displays the three orthogonal views of the currently selected correlation map in coronal, sagittal and horizontal planes. The red crosshair marks the currently selected voxel. Either click or click and drag to move to a new 3-D location. Note that this volume can be navigated in 3-D for any selected voxel location from A.

1. About/Permalink/Zoom: Clicking on “About” will bring up more information on AGEA, clicking on “Permalink” creates an URL in the address bar that effectively saves information about your current viewing state. This URL can be saved for later access to directly take you back to the current view and color scale settings. Clicking the arrows will toggle between zooming the images to fit in the window and a higher resolution, more zoomed mode. At a fixed zoom level, you may need to use the browser scroll bars to view all the images.
2. Mode Selector: Click on Correlation or Clusters to switch the viewer to different modes.
3. Position: The position of the seed selector crosshairs in millimeters in PIR (+x = posterior, +y = inferior, +z = right) orientation, where the origin is the anterior, superior, left corner of the volume.
4. Lock/Sync: The icon on the left locks the planes shown in the selected expression map B to the same position as the seed map A. Click to toggle the locking behavior. Conversely, click on the right icon to move the seed voxel to the current selected map voxel.
5. Allen Reference Atlas Label (Panel A): The structure or structural grouping from the Allen Reference Atlas to which the seed voxel (indicated by red crosshairs) belongs. Click on the name to open a window with an interactive reference atlas viewer.
6. Allen Reference Atlas Blend: This icon toggles blending Allen Reference Atlas structural delineations on the Seed Selector images for direct anatomic comparison.
7. Position: The position of the seed selector crosshairs in millimeters in PIR (+x = posterior, +y = inferior, +z = right) orientation, where the origin is the anterior, superior, left corner of the volume.
8. Allen Reference Atlas Label (Panel B): The structure or structural grouping to which the selected voxel (indicated by red crosshairs) belongs. Click on the name to open a window with an interactive reference atlas viewer.
9. Correlation: Value of the correlation at the currently selected voxel with respect to the seed voxel selected in panel A. This shows numerically how well the target voxel is correlated with the seed.
10. Gene Finder Button: (in Correlation mode only) Click this icon to find genes with enhanced expression in the voxel targeted by the crosshairs in panel A.
11. Color scale control: Use to adjust the false color mapping of the correlation map to threshold the images for regions of higher significance. All voxels with correlation within the select range are rescaled to span the color scale.
12. Download: Click to download the currently selected correlation map as raw flat file with numbers saved as floats.

‍

Using Correlation Mode

In Correlation mode, the lower panel shows orthogonal views of the selected spatial relationship map. The correlation values in the bottom figures can be interpreted as a measure of average co-expression between two voxels. The higher the correlation value between voxels, the more common it is for genes from the seed voxel to be co-expressed. Higher correlation between two voxels indicates more spatial correlation of expression and thus potentially higher possibility that the spatial regions spanned by the voxels are anatomically related. This may indicate that the voxels compared share common cell types or represent a coherent functional map. The correlation map can also be used to locate coexpressing areas in other brain regions. In the above figure, the map indicates that there is higher co-expression in thalamic structures than in other regions of the brain.

When a seed voxel is selected using the crosshairs in the images from the top panel, you can find correlations between the seed voxel and other voxels in the brain by selecting a distinct voxel in the lower panel using the crosshairs in the images from the lower panel. The correlation between the regions will be displayed below the lower panel (9).

The gene finder search facility is among the most powerful aspects of AGEA’s functionality. It enables users to search an anatomic region of interest for genes within the Allen Mouse Brain Atlas coronal database that exhibit localized enrichment. Clicking the “Find Genes” button returns a list of genes selected by the crosshairs in the upper panel (A). Voxels in the yellow to red range (top third of the range) are considered the local region of interest (target domain) while all voxels above threshold cutoff (all non-dark blue voxels) forms the contrast domain region. These domains are used in the differential search to find the genes that exhibit localized enrichment.

‍

Gene Finder Search Results

Clicking the “Find Genes” button will return a list of experiments based on the input search criteria on a separate page. The AGEA correlation map as well as the seed location in PIR orientation (in µm) is illustrated at the top of the page.

Each row of the search results includes the following information.

Column	Description
	Add the experiment to your selections by clicking the checkbox
Fold Change	A numeric value representing the fold change between expression of the target domain and the contrast domain
Experiment ID	Click the experiment number to view experimental detail page
Gene Symbol	Click Gene Symbol to return gene metadata
Gene Name	Full gene name (click to see 3-D thumbnail and perform a NeuroBlast search)
Seed Location	Image from the seed location

You will have to visually inspect the dataset to ensure the signal that AGEA detected was real and not an artifact of the image collection.

‍

Clusters Mode

Selecting Clusters mode switches the lower panels to view a data-driven hierarchical binary tree spatial organization of the brain computed from the AGEA correlation maps. To construct the decomposition, all 61,053 voxels were assigned to the root node of the tree. As we descend the tree, a node is bifurcated into two nodes to achieve maximal dissimilarity between two groups of voxels based on correlation values. The final bi-tree consists of 122,105 nodes with a maximum depth of 54 levels and 61,053 leaf nodes (one for each voxel in the brain). Effective visualization of this large data structure is via an easy-to-use Tree Depth slider mechanism to navigate the bi-tree, providing 3-D context and visualizing the multi-scale partitioning.

In the above example, when the field CA3 is selected from the top panel you can see the relationship of this region to the rest of the hippocampal formation at a tree depth of 10.

The voxels of a node are visualized with a systematic color coding scheme. All voxels of a node are assigned a color based on the ‘jet’ color scheme where the leaf node with low-order voxels are assigned shades of blue. The colors then run through green, yellow, orange, and finally, higher-order voxels are assigned shades of red.

‍