Understanding the dynamics and possibly predicting cancer progression requires knowing how evolutionary forces alter the allele frequencies in somatic cells. Population genetic studies have generated critical insights in cancer evolution (Williams et al., 2019), including two of the articles in this Research Topic.

Luddy et al. investigated the principle of evolutionary triage (Gatenby et al., 2014), which connects the frequency of a gene mutation in a population with its contribution to cancer cell fitness. A mutation that improves fitness will spur cell proliferation, increasing the likelihood of its being found in a Research Topic of tumours. Instead, a mutation that does not affect fitness will only be detected at a frequency consistent with the underlying mutation rate. Luddy et al. used this framework to investigate genes associated with tumour-immune interactions in two lung adenocarcinoma cohorts with different molecular subtypes and found distinct convergent signatures of purifying selection. Targeting genes or molecular pathways under purifying selection may have therapeutic efficacy similar to targeting driver genes and may open up new therapeutic pathways (Gatenby and Brown, 2020).

Kurpas and Kimmel developed two alternative evolution models and applied them to cancer data to test if cancer cells follow Darwinian evolution. Both models describe the combined fitness effect of driver and passenger mutations in cancer cells; the only difference is how cellular fitness affects the likelihood of occupying the space of a dying cell. In one model, population fitness is constant in the absence of mutations, consistent with Darwinian selection, whereas fitness is not constant in the alternative. The two models were fitted to breast cancer samples, and it was found that the model consistent with Darwinian selection fits the data better.

Molecular signatures can elucidate mechanisms of cancer progression and inform the clinical practice of many cancer types, perhaps most notably breast cancer (Lal et al., 2017).

Yan et al. developed a novel prognostic signature grounded on a set of necroptosis-related genes to predict breast cancer patient outcomes based on median risk scores. Patients above and below the median are classified as high-risk and low-risk, respectively. High-risk patients encounter low overall survival and worse predicted tumour, node, and metastasis stages.

Cisneros et al. studied the role of stress-induced mutagenesis (SIM) in cancer and introduced new statistical methods to identify its signature. Using somatic mutations found in tumour and normal tissue and SIM cell lines, they identified clusters of mutations consistent with stress-induced mutagenesis. These genetic clusters were less conserved in cancer, indicating a loss of regulatory control of SIM, and stress-induced mutagenesis was related to patient survival. This study shows that an evolutionary conserved adaptive mutation response, already present in bacteria, is a source of genomic instability that fuels cancer initiation, progression, and therapeutic resistance (Fitzgerald et al., 2017). This process may explain tumours’ ability to evolve under selective pressures, supporting evolutionary-based treatment strategies (Natterson-Horowitz et al., 2023).

Multi-omic studies explore the interaction between multiple biological factors to provide a more holistic view of biological processes and their relative effect on an outcome (Song et al., 2020). The development of a new multi-omics tool completes this Research Topic on cancer evolution.

Huzar et al. describe MOCA, a multi-omics toolkit for comparing gene expression and genetic evolution patterns in single cells and testing their consistency. MOCA was applied to published datasets and revealed links between genetic and phenotypic changes that can aid in understanding tumorigenesis, therapy resistance, and cancer progression. A significant shift toward single-cell technologies is expected in the coming years (Rozenblatt-Rosen et al., 2020), and MOCA will enable better exploitation of these upcoming datasets.