(2021)4:874 | Ruben Sanchez-Garcia, Josue Gomez-Blanco, Ana Cuervo, Jose Maria Carazo, Carlos Oscar S. Sorzano, Javier Vargas
DeepEMhancer is a deep learning-based approach designed to enhance cryo-EM maps, improving their interpretability and quality for atomic model building. Traditional methods, such as global B-factor correction, often fail to address the heterogeneity in map quality and can lead to undersharpening or oversharpening in different regions. DeepEMhancer, trained on pairs of experimental maps and maps sharpened using atomic models, performs automatic post-processing by applying masking-like and sharpening-like operations in a single step. Evaluations on a testing set of 20 experimental maps showed that DeepEMhancer significantly reduced noise levels and improved the resolution of the maps. The method was particularly effective in enhancing the quality of maps with heterogeneous resolution, as demonstrated in the case study of the SARS-CoV-2 RNA polymerase map. DeepEMhancer outperformed classical global B-factor-based methods and other state-of-the-art sharpening approaches, making it a valuable tool for improving the quality of cryo-EM data.DeepEMhancer is a deep learning-based approach designed to enhance cryo-EM maps, improving their interpretability and quality for atomic model building. Traditional methods, such as global B-factor correction, often fail to address the heterogeneity in map quality and can lead to undersharpening or oversharpening in different regions. DeepEMhancer, trained on pairs of experimental maps and maps sharpened using atomic models, performs automatic post-processing by applying masking-like and sharpening-like operations in a single step. Evaluations on a testing set of 20 experimental maps showed that DeepEMhancer significantly reduced noise levels and improved the resolution of the maps. The method was particularly effective in enhancing the quality of maps with heterogeneous resolution, as demonstrated in the case study of the SARS-CoV-2 RNA polymerase map. DeepEMhancer outperformed classical global B-factor-based methods and other state-of-the-art sharpening approaches, making it a valuable tool for improving the quality of cryo-EM data.