Identification of novel prognostic markers in cervical intraepithelial neoplasia using LDMAS (LOH Data Management and Analysis Software)

Abstract

BackgroundDetection of Loss of Heterozygosity (LOH) is one of the most common molecular applications in the study of human diseases, in particular cancer. The technique is commonly used to examine whether a known tumour suppressor gene is inactivated or to map unknown tumour suppressor gene(s). However, with the increasing number of samples analysed using different software, no tool is currently available to integrate and facilitate the extensive and efficient data retrieval and analyses, such as correlation of LOH data with various clinical data sets.ResultsAn algorithm to identify prognostic disease markers is devised and implemented as novel software called LDMAS. LDMAS is a software suite designed for data retrieval, management and integrated analysis of the clinico-pathological data and molecular results from independent databases. LDMAS is used in stratification of disease stages according to clinical stage or histological features and correlation of various clinico-pathological features with molecular findings to obtain relevant prognostic markers such as those used in predicting the outcome of cervical intraepithelial neoplasia (CIN). This approach lead to the identification of novel prognostic cervical cancer markers and extraction of useful clinical information such as correlation of Human Papilloma Virus (HPV) status with CIN lesions.ConclusionsA novel software called LDMAS is implemented and used to extract and identify prognostic disease markers. The software is used to successfully identify 4 novel prognostic markers that can be used to predict the outcome of CIN. LDMAS provides an essential platform for the extraction of useful information from large amount of data generated by LOH studies. LDMAS provides three unique and novel features for LOH analysis : (1) automatic extraction of relevant data from patient records and reports (2) correlation of LOH data with clinico-pathological data and (3) storage of complex data in flexible format. The first feature automates the creation of database of clinically relevant information from huge amount of data, the second feature extracts useful biomedical information such as prognostic markers in CIN and the third feature simplifies the statistical analyses of the data and allows non-statisticians to carry out the analysis. Additionally, LDMAS can be used to extract clinically useful markers from other diseases and interface to high throughput genotyping analysis software such as GDAS used to generate LOH data from Affymetrix® GeneChip Mapping arrays.