当前位置: X-MOL 学术Ultramicroscopy › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Sub-pixel electron detection using a convolutional neural network
Ultramicroscopy ( IF 2.1 ) Pub Date : 2020-11-01 , DOI: 10.1016/j.ultramic.2020.113091
J Paul van Schayck 1 , Eric van Genderen 2 , Erik Maddox 3 , Lucas Roussel 1 , Hugo Boulanger 1 , Erik Fröjdh 4 , Jan-Pieter Abrahams 5 , Peter J Peters 1 , Raimond B G Ravelli 1
Affiliation  

Modern direct electron detectors (DEDs) provided a giant leap in the use of cryogenic electron microscopy (cryo-EM) to study the structures of macromolecules and complexes thereof. However, the currently available commercial DEDs, all based on the monolithic active pixel sensor, still require relative long exposure times and their best results have only been obtained at 300 keV. There is a need for pixelated electron counting detectors that can be operated at a broader range of energies, at higher throughput and higher dynamic range. Hybrid Pixel Detectors (HPDs) of the Medipix family were reported to be unsuitable for cryo-EM at energies above 80 keV as those electrons would affect too many pixels. Here we show that the Timepix3, part of the Medipix family, can be used for cryo-EM applications at higher energies. We tested Timepix3 detectors on a 200 keV FEI Tecnai Arctica microscope and a 300 keV FEI Tecnai G2 Polara microscope. A correction method was developed to correct for per-pixel differences in output. Timepix3 data were simulated for individual electron events using the package Geant4Medipix. Global statistical characteristics of the simulated detector response were in good agreement with experimental results. A convolutional neural network (CNN) was trained using the simulated data to predict the incident position of the electron within a pixel cluster. After training, the CNN predicted, on average, 0.50 pixel and 0.68 pixel from the incident electron position for 200 keV and 300 keV electrons respectively. The CNN improved the MTF of experimental data at half Nyquist from 0.39 to 0.70 at 200 keV, and from 0.06 to 0.65 at 300 keV respectively. We illustrate that the useful dose-lifetime of a protein can be measured within a 1 second exposure using Timepix3.

中文翻译:

使用卷积神经网络的亚像素电子检测

现代直接电子探测器 (DED) 在使用低温电子显微镜 (cryo-EM) 研究大分子及其复合物的结构方面实现了巨大飞跃。然而,目前可用的商用 DED 均基于单片有源像素传感器,仍然需要相对较长的曝光时间,并且仅在 300 keV 下才能获得最佳效果。需要能够在更宽的能量范围、更高的吞吐量和更高的动态范围下操作的像素化电子计数检测器。据报道,Medipix 系列的混合像素探测器 (HPD) 不适用于能量高于 80 keV 的冷冻电镜,因为这些电子会影响太多像素。在这里,我们展示了 Timepix3 是 Medipix 系列的一部分,可用于更高能量的冷冻电镜应用。我们在 200 keV FEI Tecnai Arctica 显微镜和 300 keV FEI Tecnai G2 Polara 显微镜上测试了 Timepix3 探测器。开发了一种校正方法来校正输出中的每像素差异。使用包 Geant4Medipix 模拟单个电子事件的 Timepix3 数据。模拟探测器响应的全局统计特性与实验结果非常吻合。使用模拟数据训练卷积神经网络 (CNN) 以预测像素簇内电子的入射位置。训练后,CNN 分别从入射电子位置预测 200 keV 和 300 keV 电子的平均 0.50 像素和 0.68 像素。CNN 将半奈奎斯特实验数据的 MTF 分别从 200 keV 时的 0.39 提高到 0.70,以及 300 keV 时的 0.06 到 0.65。
更新日期:2020-11-01
down
wechat
bug