PeakNet: An Autonomous Bragg Peak Finder with Deep Neural Networks (2303.15301v3)
Abstract: Serial crystallography at X-ray free electron laser (XFEL) and synchrotron facilities has experienced tremendous progress in recent times enabling novel scientific investigations into macromolecular structures and molecular processes. However, these experiments generate a significant amount of data posing computational challenges in data reduction and real-time feedback. Bragg peak finding algorithm is used to identify useful images and also provide real-time feedback about hit-rate and resolution. Shot-to-shot intensity fluctuations and strong background scattering from buffer solution, injection nozzle and other shielding materials make this a time-consuming optimization problem. Here, we present PeakNet, an autonomous Bragg peak finder that utilizes deep neural networks. The development of this system 1) eliminates the need for manual algorithm parameter tuning, 2) reduces false-positive peaks by adjusting to shot-to-shot variations in strong background scattering in real-time, 3) eliminates the laborious task of manually creating bad pixel masks and the need to store these masks per event since these can be regenerated on demand. PeakNet also exhibits exceptional runtime efficiency, processing a 1920-by-1920 pixel image around 90 ms on an NVIDIA 1080 Ti GPU, with the potential for further enhancements through parallelized analysis or GPU stream processing. PeakNet is well-suited for expert-level real-time serial crystallography data analysis at high data rates.
- The Adaptive Gain Integrating Pixel Detector at the European XFEL. Journal of Synchrotron Radiation, 26(1):74–82, January 2019. ISSN 1600-5775. doi: 10.1107/S1600577518016077.
- Time-resolved protein nanocrystallography using an X-ray free-electron laser. Optics Express, 20(3):2706, January 2012. ISSN 1094-4087. doi: 10.1364/OE.20.002706.
- Cheetah : Software for high-throughput reduction and analysis of serial femtosecond X-ray diffraction data. Journal of Applied Crystallography, 47(3):1118–1131, June 2014. ISSN 1600-5767. doi: 10.1107/S1600576714007626.
- The ‘Seed-Skewness’ Method for Integration of Peaks on Imaging Plates. Journal of Applied Crystallography, 28(2):86–95, April 1995. ISSN 0021-8898. doi: 10.1107/S0021889894009696.
- Femtosecond diffractive imaging with a soft-X-ray free-electron laser. Nature Physics, 2(12):839–843, December 2006. ISSN 1745-2473, 1745-2481. doi: 10.1038/nphys461.
- Femtosecond X-ray protein nanocrystallography. Nature, 470(7332):73–77, February 2011. ISSN 0028-0836, 1476-4687. doi: 10.1038/nature09750.
- A peak-finding algorithm based on robust statistical analysis in serial crystallography. Journal of Applied Crystallography, 50(6):1705–1715, December 2017. ISSN 1600-5767. doi: 10.1107/S1600576717014340.
- Data reduction for serial crystallography using a robust peak finder. Journal of Applied Crystallography, 54(5):1360–1378, October 2021. ISSN 1600-5767. doi: 10.1107/S1600576721007317.
- Deep Residual Learning for Image Recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, Las Vegas, NV, USA, June 2016. IEEE. ISBN 978-1-4673-8851-1. doi: 10.1109/CVPR.2016.90.
- Untangling the sequence of events during the S 2 → S 3 transition in photosystem II and implications for the water oxidation mechanism. Proceedings of the National Academy of Sciences, 117(23):12624–12635, June 2020. ISSN 0027-8424, 1091-6490. doi: 10.1073/pnas.2000529117.
- A convolutional neural network-based screening tool for X-ray serial crystallography. Journal of Synchrotron Radiation, 25(3):655–670, May 2018. ISSN 1600-5775. doi: 10.1107/S1600577518004873.
- Structures of the intermediates of Kok’s photosynthetic water oxidation clock. Nature, 563(7731):421–425, November 2018. ISSN 0028-0836, 1476-4687. doi: 10.1038/s41586-018-0681-2.
- Serial time-resolved crystallography of photosystem II using a femtosecond X-ray laser. Nature, 513(7517):261–265, September 2014. ISSN 0028-0836, 1476-4687. doi: 10.1038/nature13453.
- Feature Pyramid Networks for Object Detection, April 2017.
- Focal Loss for Dense Object Detection, February 2018.
- BraggNN: Fast X-ray Bragg Peak Analysis Using Deep Learning. arXiv:2008.08198 [cs, eess], June 2021.
- A three-dimensional movie of structural changes in bacteriorhodopsin. Science, 354(6319):1552–1557, December 2016. ISSN 0036-8075, 1095-9203. doi: 10.1126/science.aah3497.
- Potential for biomolecular imaging with femtosecond X-ray pulses. Nature, 406(6797):752–757, August 2000. ISSN 0028-0836, 1476-4687. doi: 10.1038/35021099.
- Attention U-Net: Learning Where to Look for the Pancreas, May 2018.
- Femtosecond structural dynamics drives the trans/cis isomerization in photoactive yellow protein. Science, 352(6286):725–729, May 2016. ISSN 0036-8075, 1095-9203. doi: 10.1126/science.aad5081.
- U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv:1505.04597 [cs], May 2015.
- Data Analysis using Psocake at PAL-XFEL. Journal of the Korean Physical Society, 73(1):16–20, July 2018. ISSN 0374-4884, 1976-8524. doi: 10.3938/jkps.73.16.
- Serial crystallography using automated drop dispensing. Journal of Synchrotron Radiation, 28(5):1386–1392, September 2021. ISSN 1600-5775. doi: 10.1107/S1600577521006160.
- Light-induced structural changes and the site of O=O bond formation in PSII caught by XFEL. Nature, 543(7643):131–135, March 2017. ISSN 0028-0836, 1476-4687. doi: 10.1038/nature21400.
- Time-resolved studies of metalloproteins using X-ray free electron laser radiation at SACLA. Biochimica et Biophysica Acta (BBA) - General Subjects, 1864(2):129466, February 2020. ISSN 03044165. doi: 10.1016/j.bbagen.2019.129466.
- BraggNet: Integrating Bragg peaks using neural networks. Journal of Applied Crystallography, 52(4):854–863, August 2019. ISSN 1600-5767. doi: 10.1107/S1600576719008665.
- EfficientDet: Scalable and Efficient Object Detection, July 2020.
- James R. Weaver. Centrosymmetric (Cross-Symmetric) Matrices, Their Basic Properties, Eigenvalues, and Eigenvectors. The American Mathematical Monthly, 92(10):711, December 1985. ISSN 00029890. doi: 10.2307/2323222.
- CrystFEL : A software suite for snapshot serial crystallography. Journal of Applied Crystallography, 45(2):335–341, April 2012. ISSN 0021-8898. doi: 10.1107/S0021889812002312.
- Integration of single-crystal reflections using area multidetectors. Journal of Applied Crystallography, 21(5):471–478, October 1988. ISSN 0021-8898. doi: 10.1107/S0021889888005400.
- DIALS : Implementation and evaluation of a new integration package. Acta Crystallographica Section D Structural Biology, 74(2):85–97, February 2018. ISSN 2059-7983. doi: 10.1107/S2059798317017235.
- Chun Hong Yoon. Psocake: GUI for Making Data Analysis a Piece of Cake. In Handbook on Big Data and Machine Learning in the Physical Sciences, pages 169–178. World Scientific Publishing Co Pte Ltd, May 2020.
- Structure of photosystem II and substrate binding at room temperature. Nature, 540(7633):453–457, December 2016. ISSN 0028-0836, 1476-4687. doi: 10.1038/nature20161.