Yingjie Yu

Current development in a deep neural network has given an opportunity to a novel framework for the reconstruction of a holographic image and a phase recovery method with real-time performance. There are many deep learning-based techniques that have been proposed for the holographic image reconstruction, but these deep learning-based methods can still lack in performance, time complexity, accuracy, and real-time performance. Due to iterative calculation, the generation of a CGH requires a long computation time. A novel deep generative adversarial network holography framework is proposed for hologram reconstruction. This novel framework consists of two phases. In phase one, we used the Fresnel-based method to make the dataset. In the second phase, we trained the raw input image and holographic label image data from phase one acquired images. Our method has the capability of the noniterative process of computer-generated holograms. The experimental results have demonstrated that the proposed method outperforms the existing methods.

Structural determination of proteins is rather challenging for proteins with molecular masses between 40 - 200 kDa. Considering that more than half of natural proteins have a molecular mass between 40 - 200 kDa1,2, a robust and high-throughput method with a nanometer resolution capability is needed. Negative staining electron microscopy is an easy, rapid, and qualitative approach which has frequently been used in research laboratories to examine protein structure and protein-protein interactions. Unfortunately, conventional NS protocols often generate structural artifacts on proteins, especially with lipoproteins that usually form presenting rouleaux artifacts. By using images of lipoproteins from cryo-electron microscopy as a standard, the key parameters in NS specimen preparation conditions were recently screened and reported as the optimized NS protocol, a modified conventional NS protocol 3. Artifacts like rouleaux can be greatly limited by OpNS, additionally providing high contrast along with reasonably high-resolution images of small and asymmetric proteins. These high-resolution and high contrast images are even favorable for an individual protein 3D reconstruction, such as a 160 kDa antibody, through the method of electron tomography4,5. Moreover, OpNS can be a high-throughput tool to examine hundreds of samples of small proteins. For example, the previously published mechanism of 53 kDa cholesteryl ester transfer protein involved the screening and imaging of hundreds of samples 6. Considering cryo-EM rarely successfully images proteins less than 200 kDa has yet to publish any study involving screening over one hundred sample conditions, it is fair to call OpNS a high-throughput method for studying small proteins. Hopefully the OpNS protocol presented here can be a useful tool to push the boundaries of EM and accelerate EM studies into small protein structure, dynamics and mechanisms.