Cosmos
Cosmos is a quarterly science magazine. We aim to inspire curiosity in ‘The Science of Everything’ and make the world of science accessible to everyone.
By Cosmos
An Australian research group has developed a technology it says can reliably analyse the proteins in cancer samples on a scale never before attempted.
And it was one of those cases where necessity was the mother of invention.
ProCan (the Australian Cancer Research Foundation International Centre for the Proteome of Human Cancer based at the Children’s Medical Research Institute in Sydney) is analysing thousands of proteins in tens of thousands of cancer samples from across the world.
The project aims to create a global public library of cancer proteomics information (a proteome is the total of the proteins in a normal tissue or a cancer) that can be used to find protein signatures that predict either patient outcome or drug response in cancer.
However, the work involves obtaining reliable data from six scientific mass spectrometers operating continuously over long periods, which can be problematic, particularly when working at volume. More than 10,000 samples were processed in the first two years.
Writing in the journal Nature Communications, the researchers describe the development of computational strategies needed to correct for variations that can arise in instruments over time.
Software engineers have written code that allows unprecedented amounts of proteomic data to be processed and data scientists have devised computational methods for ensuring data can be compared and analysed successfully at any time in the cycle.
“We now have a way of designing our research cohorts so that they can be successfully integrated, regardless of whether they have been collected at different times over the life of the ProCan project,” says senior author Qing Zhong.
Lead researcher Roger Reddel says the work is designed to help find new cancer treatments and to match patients to the most effective existing treatments.
“Currently, we have treatments that can be very effective in a subset of patients, but it is often unclear which patients will respond to a specific treatment and who will suffer the side-effects without any benefit,” he says.
“By studying the protein patterns in patients’ cancers, we expect that we will be able to help predict which of the treatments that are available right now are most likely to be effective.”