Home

NEIBLAST

EyeBrowse

Eye Disease Genes

EyeSAGE

Candidate Disease Loci

Library Comparison

Search

FTP

Tips & FAQ

EyeSAGE Data on NEIBank

Duke University Eye Center's EyeSAGE project

Bowes Rickman Laboratory Hauser Laboratory

# Unique Tags in EyeSAGE: 20738 # UniGenes identified:7301 SAGE Genie date stamp: October 15, 2007

- View Candidate Eye Disease Genes and Related EyeSAGE Tags
- Search EyeSAGE Data
- View Legend in MS Excel format
- Download the Full EyeSAGE dataset in MS Access format(.mdb file)
- View the EyeSAGE dataset in MS Excel format (.xls file)

---

Related publication:
Defining the Human Macula Transcriptome and Candidate Retinal Disease Genes Using EyeSAGE, Rickman CB, Ebright JN, Zavodni ZJ, Yu L, Wang T, Daiger SP, Wistow G, Boon K, Hauser MA, Invest Opthamol Vis Sci., 2006;47(6):2305-16 (View Paper)

---

EyeSAGE is an interactive tool and database for querying human retina and retinal pigment epithelium (RPE) gene expression that simplifies the process of extracting useful information from large gene expression datasets. For example, EyeSAGE users can compare the pattern of gene expression in the retina and RPE to multiple tissues in order to identify genes that are likely to be expressed in a single cell type such as the cone photoreceptor. In addition, the database can be used for Genomic Convergence-combining expression data with available linkage information to generate lists of candidate genes for diseases such as macular degeneration. The EyeSAGE database and its uses have been described in a publication by Bowes Rickman et al. (View Paper). Click here to search EyeSAGE data.

Gene expression in the retina and RPE/choroid was characterized using Serial Analysis of Gene Expression (SAGE; http://cgap.nci.nih.gov/SAGE). Briefly, each gene is represented by a short sequence called a tag. Several hundreds of thousands of SAGE tags sequenced from a given tissue can be examined to compare the expression levels of corresponding genes. Because of the number of transcripts sampled using this technique, these SAGE libraries constitute a "reference transcriptome" of the retina and RPE.

The data can be interrogated in several ways. For example, specific gene names can be entered into the search window. Additionally, one can display all expressed genes within specific regions of the genome, such as regions linked to retinal diseases. These gene lists can be used to select and prioritize candidate genes for further analysis.


How The Dataset Was Built

Short SAGE was performed to obtain comprehensive, genome-wide expression profiles from topographically specific (macula and mid-periphery) 4 mm diameter regions of the human retina and adjacent RPE and choroid. Because these libraries were constructed from paired 4 mm punches from 5 donor eyes, true tissue-specific differences can be detected while minimizing donor-specific background. An independent library was prepared from combined macular and peripheral tissue using a modification of the SAGE technique called longSAGE. The longer tags generated by this modified technique allow more reliable identification of the corresponding expressed genes. The EyeSAGE database contains SAGE data from several different sources:

o Nearly half a million SAGE tags from the new libraries described above (Bowes Rickman et al. View Paper)
o SAGE data from 4 previously published human posterior eye short SAGE libraries (Sharon et al 2002 PNAS 99(1), 315-320).
o SAGE data from 39 additional normal tissue SAGE libraries (available at SAGE Genie; http://cgap.nci.nih.gov/SAGE/SALL?ORG=Hs)
o The EyeSAGE database also contains microarray data for comparative purposes.

The data for each library were normalized to 200,000 tags, then integrated into the EyeSAGE database, along with: the Best Gene match [21] for each tag, the UniGene cluster assignment (if available), and the gene's position in the current human genomic sequence assembly. The entire EyeSAGE database in Access was sorted by tag number. A Microsoft Excel file was then created that contains all genes with a total normalized tag count greater than 5 in the combined 8 SAGE libraries made from posterior eye tissue.

In normalized SAGE libraries, the tag count is a direct measure of the level of expression of the corresponding gene. Therefore comparison of tag counts for a given gene across multiple libraries allows the viewer to identify genes enriched in a given tissue, such as the retina or the macula. In EyeSAGE, there are columns of tag counts for each posterior eye library. Immediately adjacent are columns indicating the sum of tags from the eye ('Eye Sum'), the brain and spinal cord ('Neural Sum'), and the rest of the body ('Body Sum'). These tags sums of expression in the eye, brain and body facilitate evaluation of differential expression within regions of the eye. For example, tags derived from a rod photoreceptor cell-specific gene like rhodopsin ((RHO) have a very high Eye Sum and little or no counts in Neural Sum or Body Sum columns. In contrast, tag counts for a ubiquitously expressed gene like clusterin (CLUL1) will be high in all 3 columns. Retinal genes that are also expressed in the brain can be detected by looking for high Eye and Neural tag count sums + low tag counts in the body sum (e.g. parvalbumin, PVALB; Hippocalcin, HPCA and Calcium binding protein 1, CABP1). We have created a set of such selection criteria for various cell types. These criteria are described in the Legend (in MS Excel Format).


Long vs short SAGE and cross validation

To date the human transcriptome has been most extensively queried and represented by standard or short SAGE (14 base tags). The recent introduction of longSAGE (21 base tags) provides a platform to validate the short SAGE profiles and to directly map the longSAGE tags to the genome. Gene expression profiles derived from both of these approaches are therefore much more informative than either profile alone. With this mind, in EyeSAGE the central retina-derived longSAGE tag gene assignments and counts are displayed with the short SAGE posterior eye tags and assignments so that these can be directly compared.