Thyroid Cancer Prediction Using Gene Expression Profile, Pharmacogenomic Variants And Quantum Image Processing In Deep Learning Platform-A Theranostic Approach

The biological bigdata analysis is always a challenge to the data scientists because of the increasing size of the data day by day and the inadequacy of binary computing environment for addressing such data. The pharmacogenomic database has been used in the present work to design and develop a predictive machine for thyroid cancer for the South Asian Population based upon deep `Convolution Neural Networking (CNN). The individual variation to proneness of the disease as well as susceptibility towards drugs seek pharmacogenomic analysis with large population-wise datasets including genomics and epigenomics. In such cases, a quantum computing platform with maximum number of qubits has been suggested to make the calculations fast and accurate. The Gene expression profiling corresponding to thyroid cancer mutations have been carried out to identify the most expressed genes. The `gene expression profile pattern' has been used as the major attribute to the machine learning process. The predictions have been made from 10068 microscopic images of thyroid cancer samples collected from the South Asian Populations and using a `2 - qubits quantum computing' platform. With this small database consisting of only 10068 images, an appreciably high prediction accuracy has been achieved.