Tumor-associated antigens (TAA) have played a vital role in both diagnosis and treatment of human carcinomas, such as PSA in the diagnosis of prostate cancer and NY-ESO-1 in the treatment of liver cancer. Despite of this, the process of TAA identification has often been hampered by the complicated and laborsome lab procedures. Recently, immunoinformatics has emerged as an efficient way for TAA identification. Quite a few successful reports of novel tumor antigens identified through expression database mining can be found in the literature. Our lab has been working in this field for many years.
Traditionally, it was thought that different platforms can not be integrated together due to the difficulty of normalization. As individual series of expression data can be used separately in the case of tumor antigen identification, we believe that all kinds of expression platforms can be integrated together by gathering all the individual results. Our own experience shows that platform integration brings great affectivity for novel TAA identification.
In order to accelerate the process of tumor antigen discovery, we are now efforting to generate a publicly available Human Potential Tumor Associated Antigen database (HPtaa) with pTAAs identified by insilico computing. This database utilizes expression data from various expression platforms, including carefully chosen publicly available microarray expression data, GEO SAGE data, Unigene expression data. Please Click here for more details about source datasets. In addition, other relevant databases required for TAA discovery such as CGAP, CCDS, gene ontology database etc, were also incorporated.
In order to integrate different expression platforms together, various strategies and algorisms have been developed. Known tumor antigens are gathered from literatures and serve as training sets. A total tumor specificity penalty was computed from positive clue penalty for differential expression in human cancers, the corresponding differential ratio, and normal tissue restriction penalty for each gene.
Totally 3518 potential targets have been included in the database,
which is freely available to academic users. It successfully
screened out 41 of 82 known Cancer-Testis antigens, 6 of 18
differentiation antigen, 2 of 2 oncofetal antigen, and 7 of 12 FDA
approved cancer markers that have Gene ID, therefore will provide a
good platform for identification of cancer target genes.
Despite of the laborious work, we hope this database could help with the process of cancer immunome identification, thus help with improving the diagnosis and treatment of human carcinomas. For more information about tumor antigens and the relationship between tumor specificity and immunogenicity, please click here. For detailed explanation about how to read the results and how to make your choice, please click here.
Users are requested to cite HPtaa
home page URL,
http://www.hptaa.org and the following paper: HPtaa
database-potential target genes for clinical diagnosis and
immunotherapy of human carcinoma. Wang XS, Zhao HT, Xu QW, et al.
Nucleic Acids Res. 2006 Jan 1: 34 (Database issue):D607-12.