Combining Genomic and clinical data using bioinformatics to assist cancer clinicians and clinical cancer researchers

Abstract

The rapid advancement in genomic technologies such as DNA and RNA sequencing has greatly increased our understanding of cancer, but the implementation of genomic information in the clinic has been slow. Simultaneously, linking genomic data with clinical and pathological information in cancer research has become progressively more complex. An especially complex yet increasingly frequent task for New Zealand cancer researchers is linking their genomic data with clinical data and with data from overseas research groups. This project aims to facilitate the use of New Zealand cancer genomic data in the clinic and in research. We aimed to generate an expandable framework that: (i) combines clinical and multimodal genomic information, (ii) can contain both national and international information, (iii) allows researchers to query and link different aspects of the data to answer important clinical and genomic questions that might help clinical researchers, (iv) makes the data available on a webpage to assist clinical researchers test hypotheses and (v) in the future can assist clinicians make informed treatment decisions. This framework is designed for use with any tumour. However, the database has been initially developed using colon adenocarcinoma as an example. Potential uses of our expandable framework were reviewed with clinical researchers, clinicians and with a commercial database visualisation company. The multimodal colon adenocarcinoma data for the framework was obtained from TCGA (The Cancer Genome Atlas). A database schema was devised to make certain ‘use cases’ (extracting information to help answer questions about the clinic-pathological and genomics of the dataset) possible and a MySQL database with the appropriate tables and links was generated. RMySQL was used to wrap MySQL queries to answer important questions using the database. The results of commonly used queries were represented on a webpage created using the R shiny server software, while more complex queries were piped into further statistical analysis and visualisation using R itself. This project aims to allow researchers to integrate their own data into the existing dataset and thereafter visualise their patient relative to the pool of data available from international and New Zealand groups/organisations Keywords: colon adenocarcinoma, multimodal, personalised medicine