Back

In Search of the Perfect Model: How Cancer Cell Lines Relate to Native Cancers

Paloots, R.; Yang, Z.; Baudis, M.

2025-06-03 cancer biology
10.1101/2024.05.15.594310 bioRxiv
Show abstract

Cancer cell lines are frequently used in biological and translational research to study cellular mechanisms and explore treatment options. However, cancer cell lines may display mutational profiles divergent from native cancers or may be misidentified or contaminated. We explored how similar cancer cell lines are to native cancers to find the most suitable representations for the corresponding diseases by utilising large collections of copy number variation (CNV) profiles and applied machine learning (ML) algorithms to predict cell line classifications. Our results confirm that cancer cell lines indeed accumulate more mutations compared to native cancers but retain similar CNV profiles. We demonstrate that many relevant oncogenes and tumor suppressor genes are altered by CNV events in both cancers and their corresponding cell lines. Based on the similarities between the two groups and the predictions of the ML model, we provide some recommendations about cell lines with good potential to represent selected cancer types in in vitro studies.

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.