Relion-2.1单粒子处理教程：详细的安装与使用指南

1星需积分: 50 10 浏览量更新于2024-07-19 1 收藏 419KB PDF 举报

"Relion安装文档——Single-particle processing in relion-2.1" 这篇文档是关于"Relion-2.1"的安装教程，特别针对生物电镜领域的单粒子处理流程进行了详细介绍。Relion是一款在冷冻电子显微镜(cryo-EM)结构解析中广泛使用的软件工具，它能帮助研究人员进行高分辨率的三维重构。作者Sjors H.W. Scheres提供了这个教程，旨在引导用户完成从原始数据处理到最终结构解析的全过程。教程内容涵盖了以下几个关键步骤： 1. **束诱导运动校正**: 由于电子束对样品的影响，图像会出现运动模糊，Relion提供了束诱导运动校正功能来修复这一问题。 2. **CTF（康普顿散射函数）估计**: CTF是冷冻电镜图像质量的重要因素，通过对图像的CTF参数进行准确估算，可以提高后续分析的精确度。 3. **自动粒子挑选**: 自动化的方法可以从大量的图像中识别并提取出有价值的粒子图像，这是单粒子分析的基础。 4. **粒子提取**: 提取出的粒子图像需要进一步处理，包括定位、旋转和平移等操作，以便进行下一步的分类和重建。 5. **二维分类平均**: 通过将相似的粒子图像分组并平均，可以得到初步的结构模型，用于指导后续的三维分析。 6. **初始模型生成**: 基于二维分类的结果，可以构建初步的三维模型。 7. **三维分类**: 进一步细化粒子群，将粒子分为不同的类别，每个类别的结构可能有所不同。 8. **高分辨率三维精修**: 在获得初步模型后，Relion可以进行多次迭代优化，提升结构的分辨率。 9. **直接电子探测器的电影处理**: 直接电子探测器产生的数据具有特殊的处理需求，Relion对此提供了支持。 10. **最终地图锐化和局部分辨率估计**: 最后的处理步骤包括增强图像对比度和评估结构在不同区域的分辨率。通过仔细学习这个教程，用户预计在一天内（如果拥有合适的GPU或使用预计算结果）就能掌握Relion的使用。对于Relion的任何疑问，首先可以查阅官方维基百科的FAQ和CCPEM邮件列表的存档，如果仍有问题，可以通过订阅CCPEM邮件列表并通过指定邮箱提问。值得注意的是，不应直接向Sjors发送电子邮件，因为无法保证得到回应。本教程内容丰富，详细阐述了Relion在生物电镜单粒子处理中的应用，是科研工作者深入了解和运用Relion的强大工具。

Finally, there is an option to combine iterations through disc. If set to

Yes, at the end of every iteration all MPI slaves will write out a large ﬁle with

their accumulated results. The MPI master will read in all these ﬁles, combine

them all, and write out a new ﬁle with the combined results. All MPI salves

will then read in the combined results. This reduces heavy load on the network,

but increases load on the disc I/O. This will aﬀect the time it takes between

the progress-bar in the expectation step reaching its end (the mouse gets to the

cheese) and the start of the ensuing maximisation step. It will depend on your

system setup which is most eﬃcient. This option was originally implemented to

circumvent bugs on the network cards on our old cluster at LMB. Nowadays,

we prefer not to use this option, as it tends to be very slow when reﬁnements

reached high resolutions.

2.3 On-the-ﬂy processing: running scheduled jobs

The

Schedule button can be used to prepare jobs for future execution. Because

expected output ﬁles of scheduled jobs are already available for input into other

jobs from their

Browse buttons, it is possible to build a “pipeline” of multiple,

consecutive jobs in the

Scheduled jobs list. One can execute the scheduled

jobs using the Autorun menu on the top left of the GUI. The program will

ask you how many times you want to run the scheduled jobs. Providing a

value larger than one, will result in an iterative execution of all the scheduled

jobs. If you choose to do this, there will be another question of how many

minutes should at least pass between each repetition of all scheduled jobs. (If

the execution of all scheduled jobs takes less than this time, the program will

wait with executing the next round of scheduled jobs.) The following jobtypes

will only process micrographs that were not ﬁnished yet in a previous iteration:

Import , Motion correction , CTF estimation , Auto-picking , Particle extraction

and Movie reﬁnement . All other jobtypes will restart their calculation from

scratch again after each iteration of executing the scheduled jobs.

This setup allows convenient on-the-ﬂy processing of micrographs, as they are

being collected. One can set up a script that copies micrographs, or micrograph

movies from the microscope computer onto a processing computer (preferably

with one or more suitable GPUs for fast processing). (Because copying large

ﬁles may take some time, and you do not want to import micrographs/movies

that have not yet ﬁnished copying, this script could copy the ﬁles over with

a temporary ﬁlename, and only once they have arrived entirely could the ﬁle

then be renamed to their ﬁnal ﬁlename.) A list of scheduled jobs starting with

for example an

Import of all Micrographs/*.mrcs ﬁles, could be followed by

Motion correction , CTF estimation , Auto-picking (provided you already have

suitable templates, otherwise someone would have to manually pick some par-

ticles after a few micrographs have been acquired),

Particle extraction and then

2D classiﬁcation . This list of scheduled jobs could be repeated many times, per-

haps with at least 30 or 60 minutes between each iteration. The list of imported

micrographs will grow ever longer, and each iteration of running the scheduled

jobs, Motion correction , CTF estimation , Auto-picking and Particle extraction

will only be run on the new micrographs, while the 2D classiﬁcation will be re-

run from scratch with the total number of particles in that iteration. Inspecting

2D class averages already during data collection at the microscope may be use-

ful in order to decide whether the data is worth acquiring in the ﬁrst place, or

whether the orientational distribution is OK or one perhaps needs to collect in

regions on the grid with thicker ice. If one has a good 3D reference already,

one may even perform

3D classiﬁcation or 3D auto-reﬁne on-the-ﬂy. This may

be useful, for example, to conﬁrm whether a certain cofactor is bound to the

complex of interest.

Please note that the subset selection can be scheduled, but the user will be

prompted for input, for example by displaying the _model.star from a 2D classiﬁcation

to select suitable classes. The execution of all scheduled jobs will not continue

until the user saves this selection. Automation of this step may happen in a

future release of relion.

2.4 Helical reconstruction

Shaoda He, a PhD-student in the Scheres group, has implemented a work-

ﬂow for the processing of helical assemblies. This involves additional tabs to

the parameter-panels of the

Auto-picking , Particle extraction , 2D classiﬁcation ,

3D classiﬁcation , 3D auto-reﬁne , Particle polishing and Mask create job-types.

We do not have a separate tutorial for processing helical assemblies. The gen-

eral principles remain the same as for single-particle analysis, which is covered in

this tutorial. Therefore, users intending to use relion for helical processing are

still encouraged to do this tutorial ﬁrst. For a detailed description of the helical

options, the user is referred to the corresponding pages on the

relion Wiki, or

to Shaoda’s paper[5].

2.5 Sub-tomogram averaging

For sub-tomogram averaging, which was implemented with help from Tanmay

Bharat, a former postdoc in the Lowe group at MRC-LMB, the same holds as for

helical processing. Many general concepts remain the same as for single-particle

analysis, and users intending to perform sub-tomogram averaging in relion are

encouraged to go through this tutorial ﬁrst. For a detailed description of the

sub-tomogram averaging procedures, the user is referred to the corresponding

pages on the

relion Wiki, or to Tanmay’s paper[1].

3 Preprocessing

Although, in principle, relion can use particles that have been extracted by a

diﬀerent program, this is NOT the recommended procedure. Many programs

change the particles themselves, e.g. through phase ﬂipping, band-pass or

Wiener ﬁltering, masking etc. All these are sub-optimal for subsequent use

in relion. Moreover, gathering all the required metadata into a correctly for-

matted relion-type star ﬁle may be prone to errors. Because re-extracting

your particles in relion is straightforward and very fast, the procedure outlined

below is often a much easier (and better) route into relion.

Also, several implementations of wrappers around relion have now been re-

ported (e.g. in eman2, scipion and appion). Although we try to be helpful

when others write these wrappers, we have absolutely no control over them and

do not know whether their ﬁnal product uses relion in the best way. Therefore,

in case of any doubt regarding results obtained with these wrappers, we would

recommend following the procedures outlined in this tutorial.

3.1 Getting organised

We recommend to create a single directory per project, i.e. per structure you

want to determine. We’ll call this the project directory. It is important to

always launch the RELION graphical user-interface (GUI) from the

project directory. Inside the project directory you should make a separate

directory to store all your raw micrographs or micrograph movies in MRC for-

mat. We like to call this directory Micrographs/ if all micrographs are in one

directory, or for example Micrographs/15jan16/ and Micrographs/23jan16/

if they are in diﬀerent directories (e.g. because the micrographs were collected

on diﬀerent dates). If for some reason you do not want to place your micro-

graphs inside the relion project directory, then inside the project directory

you can also make a symbolic link to the directory where your micrographs are

stored.

Single-image micrographs should have a .mrc extension, movies should have a

.mrcs extension. If you collected movies instead of single-image micrographs,

then you can use those directly. If you calculated averages of the movies in a

diﬀerent program and you want to use both the average micrographs and the

movies, then your movies should be named in a ﬁxed manner with the same

rootname as the corresponding average micrograph, followed by an “

” charac-

ter, a “movie rootname” and a .mrcs extension. For example, if the average

micrograph is called mic001.mrc then the corresponding movie should be called

mic001_movie.mrcs (where the movie rootname would be movie).

When you unpacked the tutorial test data, the Project directory (betagal/),

and the corresponding betagal/Micrographs directory with all the recorded

剩余56页未读，继续阅读

John_Times

粉丝: 3
资源: 6

Relion-2.1单粒子处理教程：详细的安装与使用指南

relion:低温电子显微镜图像处理软件

Cryo-EM 图像模拟器：根据原子坐标创建密度图，自动包括溶剂的密度。-matlab开发

欧拉公式求圆周率的matlab代码-Relion:使用Relion软件包进行单粒子冷冻电子显微镜（cryo-EM）的脚本

ABB RELION 变电站合并单元 SMU615 操作手册

Relion 650系列线路距离保护装置REL650应用手册

Python库relion-0.4.1的安装与使用指南

OpenMPI与GPU软件安装指南：relion, FFTW, Coot等

Python脚本处理Relion文本文件指南

ABB RELION 630馈线保护测控装置产品详解与应用

Python库relion-0.2.0下载：解压缩即可使用

最新资源