Cell-free DNA is a non-invasive and accurate tool for cancer detection and monitoring. However, a current challenge is balancing the sensitivity of circulating tumour DNA (ctDNA) detection tests, with the cost of performing those tests.

A promising new area of research involves the analysis of ctDNA fragmentation patterns, which is based on the patterns of DNA degradation in the bloodstream. Nucleosome-associated DNA is protected from fragmentation, leading to a characteristic fragment length of approximately 167 base pairs. Notably, cancer patients tend to display shorter and more varied fragment lengths compared to healthy individuals.

These phenomena allow us to gain insight into the chromatin structure of the cell from which the DNA was derived, enabling assessment of disease status using low coverage sequencing data.

While previous studies have successfully distinguished between colorectal cancer (CRC) patients and healthy individuals using whole genome fragmentation analysis, this approach is too costly for a nationwide screening approach.

In this research, we employed bioinformatic techniques to understand and investigate fragmentation and its implementation as a screening tool, aiming to identify regions with distinct fragmentation patterns in colorectal cancer. Initially, we analysed important genomic features including promoters, exons and 5’UTRs to further understand the impact of these features on fragment lengths.

Next, we employed ichorCNA as a tool to measure copy number alterations in cfDNA and discovered a strong correlation between fragmentation patterns and copy number alterations. Leveraging this information, we identified specific genes exhibiting fragmentation patterns consistent with ichorCNA's estimation of tumour fraction. By calculating the average number of short fragments in these genes for a separate group of patients, we were able to correctly identify 88% of Stage 3 CRC patients, with only a 9% misclassification rate for healthy individuals.

Finally, an investigation of fragment end motifs revealed a significant enrichment of cancer-related transcription factor binding sites in CRC patients. These results represent a substantial advancement towards the development of a fragmentation-based test with practical applications as a screening tool.

Zoom details

• Zoom link: https://otago.zoom.us/j/97756704741?pwd=dHJNSW00dkFSdTk2VW1oSVlqOEJsUT09
• Meeting ID: 977 5670 4741
• Password: bioc