Cancer Prediction Using Genomic Sequences

Cancer Prediction Using Genomic Sequences

Author Name : Shraddha Sadugol, Smrithi M Menon, Namitha Manoj Pillai, Nimrita Koul

ABSTRACT

Cancer is identified as a diverse condition of several various subtypes. The Cancer Genome Atlas (TCGA) is the largest and most comprehensive collection of genetic data in the world.

We aim to build a method that can predict and classify pan-cancer tumors based on genes whose pattern of expression can distinguish between various tumor types characterized in the TCGA dataset. We analyze the RNA sequence expression data.

Most of the existing approaches are not being developed from a pan-cancer approach, and they focus on alterations on a single genetic level. We are thus training a deep neural network architecture, and on a subset of eleven thousand gene-expression samples from thirty-three tumor types, survival predictions were made. The results suggest how to apply effective treatments in one form of cancer to cancers with roughly similar genetic materials.

This is effective in reducing the mortality rate caused by cancer. Most tumors go undiagnosed until later stages. Early-stage cancer patients have a greater chance of being cured, recovering fully, and enhancing their standard of health.

Keywords— cancer, neural network, genes, RNA sequence expression data. machine learning, TCGA dataset.