Researchers from the University of Cambridge have analysed the whole genomes of approximately 12,000 tumours, originating from a variety of different tissues, along with normal samples from the same patients. The researchers were looking for patterns in the DNA of the tumours, present across samples from multiple patients (and absent from the normal samples). These patterns, known as mutational signatures, can give clues as to the biological mechanisms which led to the cancers developing as well as help to devise appropriate treatment strategies.
Principal author Serena Nik-Zainal, professor of genomic medicine and bioinformatics at the University of Cambridge, explained: 'It's like looking at a very busy beach with thousands of footprints in the sand. To the untrained eye, the footprints appear to be random and meaningless. But if you are able to study them closely, you can learn a lot about what's been going on, distinguish between animal and human prints, whether it's an adult or child, what direction they're travelling in, etc. It's the same thing with the mutational signatures. The use of whole-genome sequencing can identify which "footprints" are relevant/important and reveal what's happened through the development of the cancer.'
The analysis was carried out on data generated by the NHS 100,000 genomes project, a project run by Genomics England to sequence the whole genomes of 100,000 patients with rare disease or cancer. Dozens of organ-specific signatures were detected – some previously known and many novel. These signatures can reveal clues about a cancer's environmental origins – for instance, UV light exposure (common in melanomas) or exposure to chemical compounds eg, smoking.
They may also reveal the occurrence of specific 'driver' mutations – mutations which cause cells to become cancerous in the first place. These include mutations in genes involved in DNA repair processes, which lead to recognisable patterns of mutation accumulation in the tumour DNA. Due to the unprecedently large amount of data analysed, the researchers were able to identify many rare signatures, present in less than one to two percent of samples. The biological mechanisms behind many of the signatures remain unknown.
The authors have developed a computational tool for identifying signatures in new patient tumours, which they have named FitMS. This is a 'tool to help scientists and clinicians identify old and new mutational signatures in cancer patients, to potentially inform cancer management more effectively' according to Professor Nik-Zainal.