Researchers are using machine learning to unlock e

A new article from the University of Helsinki, published today on Nature Communication, suggests a method to accurately analyze genomic data in archival cancer biopsies. This tool uses machine learning methods to correct damaged DNA and uncover true mutation processes in tumor samples. It helps unlock huge medicinal values ​​in millions of archival cancer samples.

Molecular diagnostics helps match the right patient to the right cancer treatment. The researchers were particularly interested in DNA profiling in clinical cancer samples.

– This invaluable source is currently not used for molecular diagnostics due to poor DNA quality. Formalin causes severe DNA damage, posing an inevitable challenge for analyzing cancer genomes in preserved tissue, says lead author Qingli Guo from the University of Helsinki.

Analysis of mutation processes in cancer genomes can aid in early detection of cancer, accurately diagnose cancer, and reveal why some cancers become resistant to treatment. The new method can dramatically accelerate the development of clinical applications that can have a direct impact on the future care of cancer patients.

The new method predicted more than 90% of cancer development processes

Lead author Qingli Guo, working closely with scientists from the Institute of Cancer Research (ICR) London and Queen Mary University of London, has developed machine learning methods, named FFPEsig, to find out exactly how formalin mutates DNA.

– Our results show that normally almost half of the cancerous processes will be missed without noise correction. However, using FFPEsig, over 90% of them were predicted accurately. Qingli said.

The cancer progresses gradually. Profiling mutational processes in longitudinal samples helps identify clinical informative predictors and diagnose each tumor stage.

– Our discovery allows the characterization of clinically relevant signatures from preserved tumor biopsies stored at room temperature for decades. With an in-depth understanding of formalin’s impact on the cancer genome, our study opens up a huge opportunity to transform signature detection assays developed using large archival samples cost-effectively.

The researchers pointed out that the method currently does not completely remove artifacts that have appeared in FFPE samples showing batch effects, and the performance of the tool varies by cancer type, so care must be taken in interpreting the results. . We are also interested in further applying their methods to a much wider range of archival samples in the future.

The research was funded by Cancer Research UK, the University of Helsinki and in part by the Academy of Finland. This project is co-led by lead authors Prof. Ville Mustonen (University of Helsinki) and Prof. Trevor Graham (ICR).

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.

Sherry J. Basler