VP-Detector: A 3D multi-scale dense convolutional neural network for macromolecule localization and classification in cryo-electron tomograms
Introduction
Cryo-electron tomography (cryo-ET) is a promising imaging technique that allows three-dimensional (3D) visualization of protein or macromolecular complexes at molecular resolution in their native context. The tomographic reconstruction is generated from a tilt series imaged by rotating the biological sample. However, the resolution of tomographic reconstructions is limited by two factors: the low signal-to-noise ratio (SNR) due to the low electron dose and the missing wedge caused by the missing tilt angles. When numerous noisy copies of a complex of interest appear in a set of tomograms, subtomogram averaging (STA) can significantly increase their resolution by extracting, classifying, aligning, and averaging subtomograms in their native biological surroundings [17], [35], [36]. STA provides biological insights into the interaction and function of structures imaged under close-to-life conditions. Accurate subtomogram localization and classification, as the first two steps in the STA process, are critical for the subsequent subtomogram alignment and averaging to improve the final resolution of subtomograms. Nevertheless, extracting and distinguishing macromolecules from the complex and crowded native environment is challenging because of the low-resolution tomographic reconstruction and varied shapes and sizes of multiple macromolecules.
Traditional particle identification approaches such as template matching are employed to localize each type of macromolecule independently in tomographic reconstructions. However, due to the low SNR and missing wedge, template matching is only suitable for large complexes and ineffective for identifying small structures or macromolecular species with different states. Recently, the deep learning technique has gained revolutionary success in the fields of computer vision and image processing. In particular, convolutional neuron networks (CNNs) have been applied to object detection [24], classification [5], [10], [25], [28] and segmentation [8], [27], [31] in cryo-ET due to their natural ability to handle high noise data and achieve faster speed than template matching. However, the previously mentioned approaches for localizing and classifying subtomograms have the following deficiencies: (1) Downscaling and upscaling operations employed in CNNs can cause resolution loss, which makes particle detection in low SNR cryo-ET images difficult. (2) It is difficult to obtain adequate training data because collecting a large amount of high-quality cryo-ET images and labeling the ground truth are both time-consuming. However, these networks are typically designed with multiple trainable parameter which require large numbers of training data to prevent overfitting. (3) The number of distinct macromolecules varies significantly, resulting in the class imbalance problem that has been overlooked in previous studies.
To overcome the above limitations, we propose an efficient and automatic two-stage particle detection method called VP-Detector for localizing and classifying macromolecules in cryo-electron tomograms. VP-Detector uses a 3D multiscale dense convolutional neural network (3D MSDNet), taking advantage of the following strategies: (1) To avoid the loss of the resolution caused by the scaling operations in traditional CNNs, the 3D hybrid dilated convolution (3D HDC) module is proposed in our network, which can capture multiscale features without sacrificing resolution. (2) To train on small datasets without overfitting, 3D dense connectivity is proposed in our network, which encourages the reuse of feature maps to reduce the number of trainable parameters. (3) To address the class imbalance problem caused by the varying size of multiple macromolecules, the weighted focal loss is proposed in our network, which can reweight the loss of different classes based on prior knowledge and focus on difficult samples and rare classes.
In the experiments, VP-Detector is evaluated on two simulated datasets and one real-world dataset, and the results demonstrates that VP-Detector is accurate and effective. First, particle localization is assessed on precision, recall, miss rate, and F1-score metrics. The result shows that VP-Detector outperforms the state-of-the-art methods on the simulated tomograms with the highest precision and F1-score. Next, VP-Detector can replace manual particle picking approach on real-world tomograms. Then, the performance of particle classification is measured based on class accuracy and group accuracy. The results show that VP-Detector performs well in all groups of proteins. Finally, ablation studies are conducted to prove the effectiveness of 3D HDC, 3D dense connectivity, weighted focal loss, and training on small training sets.
Section snippets
Traditional particle detection
Over the past few years, several software packages have been designed for cryo-ET together with STA, such as Eman2 [7], RELION [2], Dynamo [4], and emClarity [15]. Most software packages support manual picking of particles on two-dimensional (2D) orthogonal slices. However, manual picking is labour intensive and subjective. Automatic picking methods are more efficient than manual picking for dealing with thousands of macromolecules in cryo-ET images. There are two types of automatic picking
Methods
In this section, we will start with the overall workflow of the two-stage VP-Detector algorithm for particle localization and particle classification. Then, we will describe the architecture of the 3D multiscale dense convolutional neural network (3D MSDNet) used in the VP-Detector, including three key contributions. First, to detect particles in low SNR cryo-ET images, the 3D HDC module is proposed in our network. Second, to train on a limited number of cryo-ET images, the 3D dense
Dataset
To verify the effectiveness of our VP-Detector for biological particle localization and classification in cryo-electron tomograms, we carried out experiments on two simulated datasets and one real-world dataset.
Simulated datasets Two simulated datasets are from the Shape Retrieval Challenge (SHREC) in 2019 and 2020 at the workshop on 3D object retrieval [11], [12]. Each dataset consists of nine sets of tomograms with 1 resolution. Only one tomogram without the ground truth
Conclusion
In this paper, we present an automatic and accurate learning-based approach, VP-Detector, for 3D bioparticle picking in cryo-electron tomograms. Our method consists of a particle localization stage and particle classification stage. The proposed network employs the 3D HDC to avoid the resolution loss, the 3D dense connectivity to encourage the reuse of feature maps to reduce trainable parameters, and the weighted focal loss to relieve the class imbalance. The support of training on small
Statements of ethical approval
The data used in this publication has been publicly available.
Declaration of Competing Interest
Authors declare that they have no conflict of interest.
Acknowledgments
The research is supported by the National Key Research and Development Program of China (Nos. 2021YFF0704300, 2017YFA0504702), the Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDA16021400 ), and the NSFC projects grants (61932018, 62072441, 62072280 and 62072283).
References (44)
- et al.
New software tools in Eman2 inspired by emdatabank map challenge
J. Struct. Biol.
(2018) - et al.
Dynamo catalogue: geometrical tools and data management for particle picking in subtomogram averaging of cryo-electron tomograms
J. Struct. Biol.
(2017) - et al.
Shrec 2020: classification in cryo-electron tomograms
Comput. Graph.
(2020) - et al.
Autom: a novel automatic platform for electron tomography reconstruction
J. Struct. Biol.
(2017) - et al.
Computer visualization of three-dimensional image data using IMOD
J. Struct. Biol.
(1996) - et al.
Deeppicker: a deep learning approach for fully automated particle picking in cryo-em
J. Struct. Biol.
(2016) - et al.
Understanding convolution for semantic segmentation
2018 IEEE Winter Conference on Applications of Computer Vision (WACV)
(2018) - et al.
Dilated convolution neural network with Leakyrelu for environmental sound classification
2017 22nd International Conference on Digital Signal Processing (DSP)
(2017) - et al.
Resolving macromolecular structures from electron cryo-tomography data using subtomogram averaging in RELION
Nat. Protoc.
(2016) - et al.
Toward detecting and identifying macromolecules in a cellular context: template matching applied to electron tomograms
Proc. Natl. Acad. Sci.
(2000)
Improved deep learning-based macromolecules structure classification from electron cryo-tomograms
Mach. Vis. Appl.
A complete data processing workflow for cryo-et and subtomogram averaging
Nat. Methods
Convolutional neural networks for automated annotation of cellular cryo-electron tomograms
Nat. Methods
Phase-aware speech enhancement with deep complex U-Net
International Conference on Learning Representations
Macromolecules structural classification with a 3Ddilated dense network in cryo-electron tomography
IEEE/ACM Trans. Comput. Biol. Bioinf.
Classification in cryo-electron tomograms
SHREC’19 Track
Deep residual learning for image recognition
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Emclarity: software for high-resolution cryo-electron tomography and subtomogram averaging
Nat. Methods
Densely connected convolutional networks
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Subtomogram averaging of COPII assemblies reveals how coat organization dictates membrane shape
Nat. Commun.
Singing voice separation with deep u-net convolutional networks
18th international Society for Music Information Retrieval Conference
Cited by (7)
Computational Methods Toward Unbiased Pattern Mining and Structure Determination in Cryo-Electron Tomography Data
2023, Journal of Molecular BiologyComputational methods for three-dimensional electron microscopy (3DEM)
2022, Computer Methods and Programs in BiomedicineComputational methods for in situ structural studies with cryogenic electron tomography
2023, Frontiers in Cellular and Infection Microbiology