Home Page of CancerSP

CancerSP is a web bench developed for analyzing genomics data and for predicting Early and Late stage of cancer using genomics data. The models used in this study were trained on TCGA genomics Level 3 data for cancer patients of six types of cancers

Major Features

• We have attempted to classify early and late stages of six cancer types of TCGA, using gene expression data. In most of the cases, Random Forest can classify early and late stage patients using gene expression data (RSEM values). We are also able to reduce the feature space from nearly 17000 genes to less than 100 genes, which could delineate early and late stages. This type of machine-learning application on high-throughput data is important to understand the mechanisms responsible for metastasis in various cancers.
• Genomics Based Prediction: This module allow users to predict cancer status(Early and Late) of a cancer patient from its genomics expression data. In this case user need to provide RSEM values of paricular genes that includes its HGNC gene symbol, and its respective RSEM value.