A study published in Nature Communications presents a novel method for detecting cellular senescence using machine learning algorithms based on nuclear features. The research team developed classifiers that accurately predict senescence in various cell types and tissues, including cancer cells and normal cells. These classifiers were validated using a range of experimental models, including mouse models of liver cancer, aging, and fibrosis, as well as patient samples with fatty liver disease. The study highlights the potential of these classifiers to detect senescence in different pathophysiological contexts and to identify and validate potential senotherapies. Senescence is a cellular response that limits the replication of old, damaged, and cancerous cells. Senescent cells undergo a stable cell cycle arrest, produce a bioactive secretome (the senescence-associated secretory phenotype or SASP), and undergo many characteristic phenotypic changes. These changes include reprogramming of metabolism, acquisition of a flat and enlarged morphology, increased lysosomal mass, rearrangement of chromatin, and nuclear changes. Senescent cells accumulate during aging, are present in cancerous and fibrotic lesions, and are often associated with disease. Research has shown that lingering senescent cells contribute to aging and disease progression. Consequently, there is growing interest in identifying drugs that selectively kill senescent cells, referred to as senolytics. Clinical trials using senolytic drugs are still in their infancy but hold enormous potential given the broad range of senescence-associated pathologies. A key requirement for the success of senolytic clinical trials is the reliable identification of senescent cells. Multiple markers to identify senescent cells exist, such as senescence-associated β-galactosidase (SA-β-gal) activity, which reflects the increased lysosomal mass of senescent cells. However, non-senescent cells such as macrophages often stain positive for SA-β-gal, and SA-β-gal can only be detected in vivo using cryosections, which complicates its use as a biomarker. Another widely used senescent marker is the cyclin-dependent kinase inhibitor p16INK4a, which is induced as part of the senescence program to arrest cells. However, p16INK4a is often deleted in cancer cells and it is difficult to detect p16INK4a in mouse tissue sections with current antibodies. Therefore, due to a combination of technical issues and the complexity and heterogeneity of senescence, there is no such thing as universal senescence markers, and there is a need to rely on multiple markers in combination. Recently, imaging-based approaches have been developed to identify senescence. While these reports prove that image-based classifiers can identify senescent cells, to what extent such classifiers can be used easily by other labs, identify a variety of senescent cell types, or be applied to other contexts and questions, is unclear. Here,A study published in Nature Communications presents a novel method for detecting cellular senescence using machine learning algorithms based on nuclear features. The research team developed classifiers that accurately predict senescence in various cell types and tissues, including cancer cells and normal cells. These classifiers were validated using a range of experimental models, including mouse models of liver cancer, aging, and fibrosis, as well as patient samples with fatty liver disease. The study highlights the potential of these classifiers to detect senescence in different pathophysiological contexts and to identify and validate potential senotherapies. Senescence is a cellular response that limits the replication of old, damaged, and cancerous cells. Senescent cells undergo a stable cell cycle arrest, produce a bioactive secretome (the senescence-associated secretory phenotype or SASP), and undergo many characteristic phenotypic changes. These changes include reprogramming of metabolism, acquisition of a flat and enlarged morphology, increased lysosomal mass, rearrangement of chromatin, and nuclear changes. Senescent cells accumulate during aging, are present in cancerous and fibrotic lesions, and are often associated with disease. Research has shown that lingering senescent cells contribute to aging and disease progression. Consequently, there is growing interest in identifying drugs that selectively kill senescent cells, referred to as senolytics. Clinical trials using senolytic drugs are still in their infancy but hold enormous potential given the broad range of senescence-associated pathologies. A key requirement for the success of senolytic clinical trials is the reliable identification of senescent cells. Multiple markers to identify senescent cells exist, such as senescence-associated β-galactosidase (SA-β-gal) activity, which reflects the increased lysosomal mass of senescent cells. However, non-senescent cells such as macrophages often stain positive for SA-β-gal, and SA-β-gal can only be detected in vivo using cryosections, which complicates its use as a biomarker. Another widely used senescent marker is the cyclin-dependent kinase inhibitor p16INK4a, which is induced as part of the senescence program to arrest cells. However, p16INK4a is often deleted in cancer cells and it is difficult to detect p16INK4a in mouse tissue sections with current antibodies. Therefore, due to a combination of technical issues and the complexity and heterogeneity of senescence, there is no such thing as universal senescence markers, and there is a need to rely on multiple markers in combination. Recently, imaging-based approaches have been developed to identify senescence. While these reports prove that image-based classifiers can identify senescent cells, to what extent such classifiers can be used easily by other labs, identify a variety of senescent cell types, or be applied to other contexts and questions, is unclear. Here,