Credit: Malvika Sudhakar
Researchers have developed a machine-learning method that can be used to identify personalised driver genes involved in starting, and accelerating the growth of breast, colon, and lung tumours1.
The method will help shortlist genes for experimental studies to better understand the progression of various types of cancers, says a team at the Indian Institute of Technology Madras in Chennai.
Existing cancer-detecting methods require a large number of cell samples to produce reliable results. In addition, they cannot be used to find personalised genes that drive tumour growth.
To overcome this drawback, the scientists designed PIVOT (Personalised Identification of driVer Oncogenes and Tumour suppressor genes), to classify genes as tumour suppressor gene (TSG), oncogene (OG) or neutral.
The researchers, led by Karthik Raman, tested the efficiency of the model in predicting driver genes such as TSGs and OGs in a large number of breast, colon and lung cancer samples. They identified new driver genes such as PRKCA, SOX9, and PSMD4 that have roles in liver, breast, skin and brain tumours.
The team also detected 1,342 unique genes for breast cancer, 1,155 unique genes for lung cancer and 1,152 unique genes for colon cancer. In a given cancer type, the number of TSGs are always greater than OGs. They also found genes that act as both TSG and OG in breast and colon cancers.
The labelling of genes as TSG or OG could potentially be used to develop personalised therapies for cancer patients, the researchers say.
