ETSAM:在冷电子断层图像中有效地分割细胞膜
Joel Selvaraj1,2, Jianlin Cheng1,2
1Department of Electrical Engineering and Computer Science, University of Missouri, Columbia, 65211, MO, United States.
bioRxiv : the preprint server for biology
|December 15, 2025
概括
一个新的AI模型,ETSAM,在冷电子断层扫描 (cryo-ET) 图像中准确地细分细胞膜. 这种先进的细分工具克服了噪音和文物挑战,改进了细胞结构分析.
科学领域:
- 细胞生物学 细胞生物学
- 结构生物学 结构生物学
- 生物物理学的生物物理.
背景情况:
- 低温电子断层扫描 (cryo-ET) 在体内可视化细胞结构.
- 对细胞膜等结构的准确细分对于理解细胞组织至关重要.
- 冷ET数据的局限性 (低SNR,文物) 阻碍了可靠的细分.
研究的目的:
- 开发一种人工智能模型,用于在冷电脑断层扫描中精确分离细胞膜.
- 为了应对噪音和冷ET数据中的工件所带来的挑战.
主要方法:
- 推出了ETSAM,这是一个基于SAM2.2的两阶段AI模型.
- 训练有素的ETSAM在83个实验式和28个模拟式冷ET断层图像的综合数据集上进行了训练.
- 在一个独立的测试套件上评估了ETSAM,该套件包括10个模拟图像和15个实验图像.
主要成果:
- ETSAM从冷ET数据中对细胞膜进行细分,实现了最先进的性能.
- 显示出高灵敏度和精度,优于其他深度学习方法.
- 与现有方法相比,实现了优越的精确召回权衡.
结论:
- ETSAM有效地在冷ET断层扫描中对细胞膜进行细分,克服了固有的数据限制.
- 该模型为分析其本地环境中的细胞结构提供了强大的解决方案.
- 由于ETSAM的开源可用性,这有助于进一步研究冷ET图像分析.
相关概念视频
Electron Microscope Tomography and Single-particle Reconstruction
2.8K
Transmission electron microscopy (TEM) can be used to determine the 3D structure of biological samples with the help of techniques such as electron microscope tomography and single-particle reconstruction. While single-particle reconstruction can examine macromolecules and macromolecular complexes in vitro conditions only, tomography permits the study of cell components or small cells in vivo.
Electron Tomography
Electron tomography can be performed either in TEM or STEM (scanning transmission...
Electron Tomography
Electron tomography can be performed either in TEM or STEM (scanning transmission...
2.8K
Cryo-electron Microscopy
4.1K
Conventional electron microscopy (EM) involves dehydration, fixation, and staining of biological samples, which distorts the native state of biological molecules and results in several artifacts. Also, the high-energy electron beam damages the sample and makes it difficult to obtain high-resolution images. These issues can be addressed using cryo-EM, which uses frozen samples and gentler electron beams. The technique was developed by Jacques Dubochet, Joachim Frank, and Richard Henderson, for...
4.1K


