About the EADHB Database
|
About the EADHB Database |
More detailed information can be found in the FAQs.
The Electronic Atlas of the Developing Human Brain (EADHB) project is a collaboration bewtween the Institute of Human Genetics in Newcastle, UK, and the MRC Human Genetics Unit in Edinburgh, UK, and is part of the NIH Human Brain Project. The database was developed as part of this project and is based on the Edinburgh Mouse Atlas gene expression database (EMAGE) and the EURexpress databases. It is designed to be an openly available resource to the research community to hold gene expression patterns during the early development of the human brain.
As in normal databases there is a good deal of text-based information (for section descriptions, probe data, where the gene is expressed etc...) but what makes it unique are the spatial 2D and 3D visualisations of the gene expression data.
The EADHB Database is built on a framework containing two parts. The first is a set of anatomical terms at various stages of development (known as an ontology). This ontology is used to define sites of gene expression using a set of standard descriptions and to link the expression data to an 'anatomical tree' . The second is a standard set of 3D models at different stages of development. Currently this set of models spans from Carnegie Stage12 to Carnegie Stage 23. The experimental data is captured and converted to digital format and then mapped to the appropriate 3D model.
As the database has both text and spatial data stored, we are then able to search the database using either words, (e.g. keywords of the anatomy tree or by gene name), or by searching a specific area of space within the model. This is discussed further in later sections in these pages.
What types of data does the EADHB Database include?
The EADHB Database, in its current state, only contains a small number of expression patterns and data entries. This number will increase as more in-situ data is collected, mapped and uploaded to the central database. Further data will be collected from our own research projects as well as from external users wishing to use the database for their own data.
The data collected so far includes both radioactive and non-radioactive in-situ hybridisation experiments (detected using anti-sense mRNA) as well as immunohistochemistry data (using DAB labelled antibodies). The original data is from paraffin embedded sections (tissue fixed in 4% PFA), digital images of the sections are captured and this data is mapped (textually and to the corresponding 3D model) in a local database, checked and then uploaded to the central, publicly acessible, database.
The data available at the moment is all single section data for a single entry, although some entries have 3D visualisations which show more than one section if more than the single entry section has been mapped. It is hoped that 3D data entry will be available in the future together with more specialised immunohistochemistry entry tables (this is discussed further in later sections of this page).
Why has the EADHB Database been developed?
One of the major reasons for developing a gene expression database of this type is to allow direct spatial comparison between two or more complex 3D expression patterns. The EADHB database is unique in that you are able to search through both the EADHB database and the EMAGE database and directly compare human and mouse data. This function not only allows the user to compare two genes in human, e.g. to aid study into any potential genetic interactions, but also allows for the comparison of a single gene expression pattern to that of its mouse homologue. It is important to realise, especially in the context of brain development, that there are many differences in the expression of particular genes between mouse models and their human counterparts.
Another important advantage of the EADHB database is that it functions to eliminate the need to describe complex 2D and 3D gene expression patterns using text alone. The anatomy ontology (anatomy trees built with lists of standardised words for naming specific regions) currently being used is that of the Edinburgh Human Developmental Anatomy Database which encompasses all developing structures from CS1 to CS20 but is not detailed for developing brain structures. This ontology is being extended and refined (by Prof Luis Puelles and his colleagues, University of Murcia,Spain; Prof Salvador Martinez, University of Miguel Hernandez, Alicante, Spain and Dr Duncan Davidson, MRC Human Genetic Unit, Edinburgh UK) and will be incorporated into the EADHB database as it is developed. At present the ontology provides a framework for text descriptions of gene expression patterns, however, there may still be problems when using words alone to describe expression patterns even though the user has knowledge of the correct anatomical terminology. In using the spatial mapping features of the database, the user simply maps the expression pattern to the standard 3D model and any anatomical domains which have been painted onto the model will automatically be specified as either having gene expression present or not.
The database also allows the user to store and retrieve experimental data and results in an orderly fashion on a local database, (eliminating the need for sorting through racks of slides each time someone needs to see a result), while actively encouraging data sharing bewteen the wider community.
Who has developed the EADHB Database?
The EADHB gene expression database has been developed as part of an NIH Human Brain Project funded collaboration between the Institute of Human Genetics in Newcastle, UK and the MRC Human Genetics Unit in Edinburgh, UK. The database itself is based upon the EMAGE mouse gene expression database, which was created by the Mouse Atlas Group at the MRC Human Genetics Unit in Edinburgh, UK as a part of the Edinburgh Mouse Atlas Project, (EMAP). The EADHB database is run through the Institute of Human Genetics, Newcastle, and is maintained by the staff at the Mouse Atlas Group, (for a list of people involved with running the programming and maintenance of the EADHB database please click here).
How is the data entered into the database?
Data is entered using a program called MAPaint, which has been developed as part of the Edinburgh Mouse Atlas group's EMAP project. The software is used to 'warp' the digitised section data to the standard 3D model. Although the data used currently is 2D section data from either tissue in situ hybridisation or immunohistochemistry experiments, it is also possible to use lateral views of wholemount data and it is hoped that in the future 3D OPT scans of expression data will be able to be incorporated into the database.
The original data is digitised and read into MAPaint along with the standard 3D model for the appropriate stage. The equivalent section in the 3D model is found to match that of the 2D section and the co-ordinates written out. A process known as 'warping' is then performed on the 2D section data. This involves 'tie-ponts' being created (effectively matching various landmark points on the 2D digitised experimental section and the equivalent section of the 3D standard model) and the section being 'warped' to fit by the use of a triangular mesh which underlies the original images.
Once the sections have been matched and the warping process performed, the expression data is then extracted from the image by a technique known as thresholding. This process effectively shows all colours or intensity levels over a certain limit set by the user. This signal data is then placed onto the 3D model in the relevant place according to the section plane chosen. By altering the threshold values the user is able to map different levels of expression showing strong, moderate and weak expression, possible and un-detected expression as well as areas that have not been examined in the experiment. Some manual editing is usually required to delete artefact from being included in the expression patterns, sometimes areas of dirt, folding or variations in background staining will show up as expression and should be removed from the final results before saving.
Once the expression patterns have been mapped and saved, the data can be read into the database submission interface along with information about the experiment, user and specimen and with images of the original data. The availability of the original section data images in the database is very important as users can examine this independently of the mapped data (where the thresholding process introduces subjective interpretation of the data). The patterns saved are then be added to the model and a screen shot of the data on the model is created for display purposes in the database and a 3D representation movie can also be created and included for users to download and view - giving a better indication of where the expression is located in relation to other data being viewed.
What types of searches can be performed?
The database has both text and spatial elements underlying it, which allows the user to perform searches using text strings as well as spatial domains.
Searches using text strings can include either the gene name of interest or the anatomical region of interest. Spatial searches can be performed using predefined anatomical regions that have been painted onto the models and linked to the relevant ontology, but the user can also perform searches by physically painting an area of the model and searching this specific area. Any genes or anatomical features linked with any of this area will then be found in the search.
The way the search engine is setup makes it very versatile. The user can search for which gene they are interested in (e.g. search for FGF8 or SHH) and the search will report which stages and which anatomical domains the gene is present in. However, the user can also specify which anatomical feature they are interested in at a specific stage and find out which genes are expressed in that feature. It is important to remember that the list will only show those results that have been included in the database and will not reference any other source. For this reason, some genes may or may not show up in searches if the expression has not been included for a specific region (even if the gene is known to be expressed there - i.e. experiments where certain areas of tissue have not been examined).
The EADHB database also includes a function to search the EMAGE database along with its own data. This means that users can search for entered human data but also be shown any mouse homologues which may be present in the EMAGE database.
How can I submit data to the EADHB Database?
Detailed instruction on how to use the EADHB database are not available on-line yet. However, the database help menu does include instructions on how to create a local copy of the database, how to input entries and how to upload to the central database. For a closer look at the submission process please see the Guidelines for Submission of Data page.
While the database is currently available for download and use, it is strictly only a prototype database and will undergo various changes in the near future.
At the moment the database can only accept a single section of data for a single specimen at a single stage. It is hoped that in the future this will be modified to enable input of 3D wholemount data as well as 3D mapped data from multiple serial sections. This will enable wider studies of gene expression boundaries which will have a knock on effect with regards to studying anatomical boundaries delineated by gene expression results. It will also allow for wholemount data captured using OPT (optical projection tomography) to be mapped in 3D rather than just a snapshot of a wholemount specimen.
We can add immunohistochemistry data into the database at the moment but the interface is mainly set up to accomodate in situ results using cRNA probes. This will be changed to allow more accurate details of immunohistochemical experiments to be added. The user will also be able to add data from laser dissection and microarray experiments, provided the location of the tissue is precisely known for anatomical mapping purposes.
As mentioned above, the anatomy ontology is being extended and refined (by Prof Luis Puelles and his colleagues, University of Murcia,Spain; Prof Salvador Martinez, University of Miguel Hernandez, Alicante, Spain and Dr Duncan Davidson, MRC Human Genetic Unit, Edinburgh UK) and will be incorporated into the EADHB database as it is developed.
While the anatomy ontology is being refined, the standardised 3D models will have more detailed anatomical domains painted onto them and linked to the anatomy tree. This will have a follow on effect in allowing the mapped gene expression data to be linked to specific anatomical features for searching and providing more accurate data analysis. This anatomy will also be painted on to a series of higher resolution Histology Models which will be completed as part of future work of the EADHB project.
Less noticeable (but still important) changes are to the help files and instruction sheets which are still largely adapted from the EMAGE and EURexpress versions of the database. These will be re-written to be more pertinent to the EADHB database and the help files more useful for problems people may encounter with regards to the use of the human database as well as instructions on how to enter the varous types of data and search the database with cross-correlation to the EMAGE database.
We welcome any feedback or suggestions you may have about future improvements or additions to the database. Please contact us at eadhb@ncl.ac.uk.