Consensus algorithm developed by Singapore scientists improves identification of treatment targets in individual cancers

Consensus algorithm developed by Singapore scientists improves identification of treatment targets in individual cancers

Scientists from Singapore’s Agency for Science, Technology
and Research (A*STAR)’s Genome Institute of Singapore (GIS), the National
Cancer Centre Singapore (NCCS) and the National University of Singapore (NUS),
have developed
a consensus algorithm
which integrates information from many different computer
programmes for accurate prediction for treatment targets in individual cancers.
The outcome was a significant improvement over the performance of any single

The work, recently published in Cancer Research,
was jointly led by Dr Denis Bertrand and Professor Niranjan Nagarajan from

The press release explains that breakthroughs in DNA
sequencing technologies have allowed researchers to determine the complete
genetic makeup of cancers. Now the challenge is to analyse the massive datasets
to understand the unique genetic basis of an individual’s disease.  Cancer cells have thousands of genetic lesions
but only a few of these mutations give rise to a tumour. Identifying the ‘driver’
mutations that promote the uncontrolled growth of cancer cells in the body is a
key challenge for precision oncology.

Researchers around the world and in Singapore are now working
to develop new computer algorithms, and participating in large collaborative
projects such as The Cancer Genome
(TGCA) to unlock the mysteries of different cancers. (TGCA is a collaboration between the US-based
National Cancer Institute (NCI)
and the National Human Genome Research Institute (NHGRI) that has generated
comprehensive, multi-dimensional maps of the key genomic changes in 33 types of
cancer. The TCGA dataset, comprising more than two petabytes of genomic
data, has been made publicly available. This genomic information helps the
cancer research community to improve the prevention, diagnosis, and treatment
of cancer.)

The Singapore researchers analysed data from more than 3,000
tumours, across 15 different cancer types including colon, breast, lung,
stomach and liver cancer and studied 18 different existing algorithms. The scientists
found that each algorithm on its own could not identify driver mutations in a
significant proportion of patients. Furthermore, no single method was able to
identify treatable drivers in more than 60 percent of patients.

But they noted that the methods had very different strengths.
By combining them, the new system, known as ConsensusDriver, was able to
identify treatment targets in nearly all patients studied, 80 percent of whom
could be treated with existing drugs. 

Workflow for the ConsensusDriver system in analysing patient tumours and identifying
target treatments
‍(Copyright: A*STAR’s Genome Institute of Singapore)

Dr Denis Bertrand, Staff Scientist at GIS and lead author of
this work said, “Developing ConsensusDriver and working with The Cancer Genome
Atlas has been an eye-opening experience. This is collaborative science on an
international scale and we are making rapid advances in being able to give the
right drug to the right patient at the right time.” 

Professor Nagarajan, Associate Director and Senior Group
Leader at GIS, noted, “It is remarkable that computer algorithms have become a
new weapon in the battle against cancer. Instead of clubbing cancer cells with
drugs indiscriminately, we are now trying to computationally pinpoint genetic
weaknesses to target them with drugs more precisely.”

GIS Executive Director, Professor Ng Huck Hui, said, “The complexity of cancer
genetics is one of the biggest challenges that we face in treating it. By
precisely identifying actionable mutations, and tailoring treatments to
individuals, we are moving a step closer to precision medicine. I am delighted
to note the ongoing development of new algorithms and technologies by GIS
scientists to achieve this vision.”