The ATM serine/threonine protein kinase, encoded by the ATM gene, plays a crucial role in regulating the cell’s mechanisms to repair DNA damage, and mutations in ATM increase the risk of cancer and impact how cancer progresses. However, many of the identified changes in ATM are currently classified as “variants of unclear significant”, i.e. their clinical relevance is not clear. Researchers have now investigated all possible variants in the coding regions of ATM with respect to their effect on cell survival. Their study, recently published in Cell, aimed at enhancing cancer risk assessment and prognosis prediction.

The researchers systematically assessed all 27,513 possible single nucleotide variants (SNVs) in ATM. For genome editing, they used prime editing, a tool which allows to precisely and specifically rewrite single base pairs in the genome while avoiding DNA double strand breaks. Additionally, they developed a highly accurate deep learning model (DeepATM) for predicting functional impacts of ATM variants. Integration of their data with the UK Biobank cohort identified 382 high-risk SNVs linked to cancer.