High Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China; Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China.
High Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China.
Comput Methods Programs Biomed. 2022 Jun;221:106871. doi: 10.1016/j.cmpb.2022.106871. Epub 2022 May 11.
Cryo-electron tomography (cryo-ET) with subtomogram averaging (STA) is indispensable when studying macromolecule structures and functions in their native environments. Due to the low signal-to-noise ratio, the missing wedge artifacts in tomographic reconstructions, and multiple macromolecules of varied shapes and sizes, macromolecule localization and classification remain challenging. To tackle this bottleneck problem for structural determination by STA, we design an accurate macromolecule localization and classification method named voxelwise particle detector (VP-Detector).
VP-Detector is a two-stage particle detection method based on a 3D multiscale dense convolutional neural network (3D MSDNet). The proposed network uses 3D hybrid dilated convolution (3D HDC) to avoid the resolution loss caused by scaling operations. Meanwhile, it uses 3D dense connectivity to encourage the reuse of feature maps to reduce trainable parameters. In addition, the weighted focal loss is proposed to focus more attention on difficult samples and rare classes, which relieves the class imbalance caused by multiple particles of various sizes. The performance of VP-Detector is evaluated on both simulated and real-world tomograms, and it shows that VP-Detector outperforms state-of-the-art methods.
The experiments show that VP-Detector outperforms the state-of-the-art methods on particle localization with an F1-score of 0.951 and a precision of 0.978. In addition, VP-Detector can replace manual particle picking in experiment on the real-world tomograms. Furthermore, it performs well in classifying large-, medium-, and small-weight proteins with accuracies of 1, 0.95, and 0.82, respectively. Finally, ablation studies demonstrate the effectiveness of 3D HDC, 3D dense connectivity, weighted focal loss, and training on small training sets.
VP-Detector can achieve high accuracy in particle detection with few trainable parameters and support training on small datasets. It can also relieve the class imbalance caused by multiple particles with various shapes and sizes.
在研究生物大分子在其天然环境中的结构和功能时，低温电子断层扫描（cryo-ET）结合亚断层平均化（STA）是必不可少的。由于信号噪声比较低、断层重建中的缺失楔形伪影以及形状和大小各异的多个生物大分子，生物大分子的定位和分类仍然具有挑战性。为了解决 STA 结构测定中的这一瓶颈问题，我们设计了一种名为体素粒子探测器（VP-Detector）的准确生物大分子定位和分类方法。
VP-Detector 是一种基于 3D 多尺度密集卷积神经网络（3D MSDNet）的两阶段粒子检测方法。所提出的网络使用 3D 混合扩张卷积（3D HDC）来避免由于缩放操作而导致的分辨率损失。同时，它使用 3D 密集连接来鼓励特征图的重用，以减少可训练参数。此外，提出了加权焦点损失，以更关注困难样本和稀有类，从而缓解了由于各种大小的多个粒子引起的类别不平衡。在模拟和真实断层图像上评估了 VP-Detector 的性能，结果表明 VP-Detector 优于最先进的方法。
实验表明，VP-Detector 在粒子定位方面的表现优于最先进的方法，F1 得分为 0.951，精度为 0.978。此外，VP-Detector 可以在真实断层图像的实验中替代手动粒子挑选。此外，它在分类大、中、小重量蛋白方面表现良好，准确率分别为 1、0.95 和 0.82。最后，消融研究证明了 3D HDC、3D 密集连接、加权焦点损失和在小训练集上训练的有效性。
VP-Detector 可以在具有少量可训练参数的情况下实现高精度的粒子检测，并支持在小数据集上进行训练。它还可以缓解由于多种形状和大小的多个粒子引起的类别不平衡。
