Cancers begin when changes in genes called mutations cause cells to start multiplying uncontrollably.
Mutations can be the result of damage to DNA caused by external agents, like UV light or chemicals in cigarettes, or caused by ‘mistakes’ when DNA is replicated in cells before they divide.
Cancer cells can have large numbers of these mutations in their DNA, and often cells within the same tumour are heterogenous, meaning that their DNA differs from cell to cell.
However, there are certain ‘types’ of DNA mistake that occur in most cancers. One of these mistakes is called aneuploidy or, more simply, chromosome copy number changes.
This means that a cell has too many, or too few, chromosomes, coiled structures of DNA that store our genetic code, compared to normal. To give an example, a human cancer cell may have 45 or 47 chromosomes as opposed to the normal 46.
Chromosome copy number changes can make the genetic code of cancer cells highly complex. Imagine trying to read a book that’s missing entire chapters, or has chapters duplicated and inserted back in the wrong place.
It results in tumours growing and spreading more quickly, and can result in resistance to treatment, meaning that the prognosis for the patient is poorer.
Seeing the bigger picture
Mutations have been studied in cancers for decades and we have been able to identify simple ones using a technique called DNA sequencing to see how the DNA differs from that of normal cells. However, the understanding of complex mutations, like aneuploidy, is still a gap in our knowledge.
“To stay one step ahead of cancer, we need to anticipate how it adapts and changes,” says Dr Nischalan Pillay, Associate Professor in Sarcoma Genomics at University College London (UCL).
“Mutations are the key drivers of cancer, but a lot of our understanding is focused on changes to individual genes in cancer. We’ve lacked the tools to better understand how vast swathes of genes can be copied, moved around or deleted without catastrophic consequences for the tumour, and how these could be used to classify cancer.”
Now, an international team of researchers at UCL and the University of California, San Diego (UC San Diego), co-led by Pillay and with funding from Cancer Grand Challenges, are using artificial intelligence to find new ways to identify and classify these complicated genetic changes.
Finding the patterns
Using machine learning techniques, the team developed an algorithm to investigate and categorise the size and scale of DNA changes across the genome, a cell’s complete genetic code, when cancer starts and grows.
This algorithm might sound complex, but it’s not too different from one you might be familiar with. The way it analyses data to find patterns is similar to how the streaming service Netflix uses data on your viewing habits to recommend other films and series to watch.
The scientists looked for common patterns in how the chromosomes are reorganised in the tumours of over 9,000 patients with 33 different types of cancer.
From this massive amount of data, they were able to identify 21 common faults to the structure and number of chromosomes in tumours and categorise them into different “genres” called copy number signatures.
These copy number signatures have provided a ‘blueprint’ of features that are common between cancer types that provides new insight into the complex genomes of cancers.
Now that these signatures have been identified, there’s the potential to use them as targets for new therapies.
“Cancer is a complex disease, but we’ve demonstrated that there are remarkable similarities in the changes to chromosomes that happen when it starts and how it grows,” says Dr Ludmil Alexandrov, Associate Professor at UC San Diego and co-lead author of the study.
“Just as Netflix can predict which shows you’ll choose to binge watch next, we believe that we will be able to predict how your cancer is likely to behave, based on the changes its genome has previously experienced.”
The team then further investigated the copy number signatures that most strongly affected outcomes for cancer patients.
Of the 21 signatures they identified, they found that tumours associated with the worst survival outcomes were ones that had gone through a process called chromothripsis, where the DNA in a chromosome completely shatters and reforms incorrectly.
For example, they found that people with glioblastoma, an aggressive type of brain tumour, had worse survival outcomes if cells in their tumour had undergone chromothripsis.
What’s next?
The team hopes that the algorithm can now be refined to enable doctors to find out how a cancer is likely to behave, based on the genetic traits it acquires when it starts and the changes it picks up as it grows.
And to accelerate progress, they’ve shared the blueprint of signatures they’ve found, and the tools they used to uncover them, freely with the global scientific community.
That way, other scientists can build their own Netflix-style libraries of chromosome changes, based on data obtained from analysing tumours in their own research.
“We want to get to the point where doctors can look at a patient’s fully sequenced tumour and match the key features of the tumour against our blueprint for genomic faults,” Alexandrov adds.
“Armed with that information, we believe that doctors will be able to offer better and more personalised cancer treatment in the future.”
Jacob