In this study, we aimed to leverage various microscopy techniques to collect a rich dataset of microscopic images, thereby validating the generalizability and practicality of our deep learning–based method. We express our deep gratitude to Novel Optics (Ningbo, China) for the invaluable opportunity to use their NCF950 fluorescence microscope. The two original fluorescence images used to create the cover image of the October 15 issue of Biophysical Journal were captured with the NCF950. In these images, the red channel represents cell adhesion and endoplasmic reticulum structures in a live COS-7 cell, labeled with paxillin-mCherry and mCherry-KDEL, and the green channel corresponds to clathrin-coated pits and microtubules, labeled with clathrin-eGFP and EMTB-3×eGFP.
This multi-structure overlay image validates the effectiveness of our deep- learning model in observing real live-cell samples. It serves as critical data for assessing the scientific validity and practical applicability of our method.
When studying the interactions between different types of subcellular structures, scientists often encounter the challenge of temporal discrepancies between channel-specific imaging, which can compromise image quality. The deep learning–based technique for simultaneous imaging of multiple subcellular structures proposed in this paper addresses this challenge, enabling true multi-structure imaging while ensuring high-quality single-structure data acquisition.
The deep learning model developed in this study provides essential technical support for researchers seeking reliable tools to advance biomedical research. It also offers valuable insights into multi-structure simultaneous imaging for those proficient in deep learning but inexperienced in biological applications. This study demonstrates the research value of interdisciplinary collaboration among biomedicine, engineering, and information technology, making it relevant to a wide audience.
More information about our work can be found at https://person.zju.edu.cn/en/yingkexu.
— Luhong Jin, Jingfang Liu, Heng Zhang, Yunqi Zhu, Haixu Yang, Jianhang Wang, Luhao Zhang, Cuifang Kuang, Baohua Ji, Ju Zhang, Xu Liu, and Yingke Xu