Scripts and sources of the block-based reconstruction/refinement.
These scripts and codes are under WTFPL verion 2 license, a GPL-campatible free license.
The modification version of RELION is under the same license of RELION, which is GPL ver.2.
Please cite our paper: NATURE COMMUNICATIONS (2018) 9:1552 or DOI: 10.1038/s41467-018-04051-9 if our block-based reconstruction idea or these codes work.
Fix a critical bug that the rlnDefocusU/V would be completely wrong.
Download and reinstall this.
https://github.com/homurachan/Block-based-recontruction/blob/master/relion-3.1.2_BBR.zip
Created a page in Wiki:
https://github.com/homurachan/Block-based-recontruction/wiki/How-to-fit-block-into-original-map
describing how to fit map of blocks after refinement into their original position.
Created 2 pages in Wiki:
https://github.com/homurachan/Block-based-recontruction/wiki
https://github.com/homurachan/Block-based-recontruction/wiki/How-to-determine-position-of-block
which describe how to combine blocks into full map and how to determine position of block using UCSF Chimera.
Uploaded a small dataset of reovirus, which can be used for practicing BBR:
https://1drv.ms/u/s!AghYYiVwSrFmhAxUTa0swcXgWJn0?e=BnW1bJ , information is inside of the compressed files. micrograph_ctf.star and refine.star are also included.
Add feature "--remove_nan" in relion_star_handler, which should be used together with "--discard_on_stats". This options read star files and remove the image item that has infinity or NaN in it. It was inspired by our colleague who suffered from a random blank-micrograph.
Move old readme.md to https://github.com/homurachan/Block-based-recontruction/blob/master/old_version/Readme.old2.md
Uploaded the modified version of RELION 3.1.2. Now you can do the process within RELION GUI and pipelines.
Just download the "relion-3.1.2_BBR.zip" (https://github.com/homurachan/Block-based-recontruction/blob/master/relion-3.1.2_BBR.zip), unzip and install like original version of RELION. It can replace RELION3.1.2 since no function change was made to rest parts of RELION. You will find the "Block-based Recons" option on the GUI.
Input particles: Just your particles.
Symmetry of particles: Any symmetry within RELION is supported.
X/Y/Z position of reconstruction: The same Point_X, Point_Y, Point_Z below.
Correct/Inverted handerness: Choose the handerness of sub-particles. If both were "NO", no defocus change will be made (for small particles).
X/Y-size of the micrograph: The image size of the micrograph. Program will drop sub-particles that the center lays outside of the micrograph.
Rounded translation: The OriginX/Y will be set to integer rather than float. Just for debug usage.
Samples of input are shown below.
After finish the job, use the output file as "Refined particles STAR file" on I/O of Particle extraction. Please set re-center refined coordinates to (0,0,0). Then adjust the particle box size to block box size.
Reconstruct the output star file by relion_reconstruct --i OUTPUT.STAR --ctf --sym c1 --subset 1/2 --o half1/half2.mrc . (Optional below) Calculated the FSC curve. You can run 'relion_star_handler --i OUTPUT.STAR --o out.star --do_invert_BBR_handerness' to invert the handerness of the star. Keep the one with higher FSC.
Just ordinary single particle analysis. Remember always use local refinement (Initial angular sampling <= Local searches from auto-sampling in 3D auto-refine)."
包含了分块重构所需的脚本及合并的小程序,遵循WTFPL verion 2 license, a GPL-campatible free license. 改编版RELION遵循GPL ver2 license。 如果这些代码或分块重构算法对你的研究有帮助，请引用: NATURE COMMUNICATIONS (2018) 9:1552 or DOI: 10.1038/s41467-018-04051-9
修正了一个严重bug，当循环时sub-particle的defocusU/V会发生完全错误。
下载并重新安装以下版本。
https://github.com/homurachan/Block-based-recontruction/blob/master/relion-3.1.2_BBR.zip
https://github.com/homurachan/Block-based-recontruction/wiki/How-to-fit-block-into-original-map
描述如何把分块重构好的分块fit进原始位置。因为多数情况下refine好的分块会有一个小旋转。
创建了两页Wiki:
https://github.com/homurachan/Block-based-recontruction/wiki
https://github.com/homurachan/Block-based-recontruction/wiki/How-to-determine-position-of-block
分别描述如何把多个重构好的分块合并成一个完整的map；以及如何用Chimera确定每个分块的中心位置。
上传了一个可以用来练习分块重构的reovirus的小数据集:
https://1drv.ms/u/s!AghYYiVwSrFmhAxUTa0swcXgWJn0?e=BnW1bJ ,包含micrograph_ctf.star和refine.star,其他信息在压缩文件的info里.
在relion-3.1.2_BBR.zip的relion_star_handler里增加了--remove_nan功能，需与--discard_on_stats连用。作用是读取star文件，把含有无穷大的图像的条目去除，不改变图像。
更新了修改版的RELION 3.1.2. 现在可以使用修改版RELION在GUI中完成分块重构.
下载"relion-3.1.2_BBR.zip" (https://github.com/homurachan/Block-based-recontruction/blob/master/relion-3.1.2_BBR.zip), 解压后比照原始版RELION安装. 该修改版可以完全替代原始版RELION3.1.2. 打开后可以在主界面看到"Block-based Recons"。
Input particles: 输入文件.
Symmetry of particles: 任何RELION支持的对称性都可以使用.
X/Y/Z position of reconstruction: 如同下文所述的Point_X, Point_Y, Point_Z.
Correct/Inverted handerness: 选择颗粒手性，如果都选NO，那么sub-particle的defocus不会改变，适用于较小的颗粒。
X/Y-size of the micrograph: micrograph的尺寸，必须是particle extraction里使用的. 程序会把中心落在micrograph外的sub-particle扔掉.
Rounded translation: OriginX/Y变为整数，只在测试时使用（防止插值误差）.
下图是一些参数样本.
生成的star文件放到relion -> Particle extraction -> I/O -> OR re-extract refined particles = yes -> "Refined particles STAR file"这里,同时选中OR: re-center refined coordinates = yes.接下来就重新选择Particle box size为block大小就行了.
判断手性,用'relion_reconstruct --i OUTPUT.STAR --ctf --sym c1 --subset 1/2 --o half1/half2.mrc'重构组合好/生成好的star文件. 如果不想判断手性也行，可忽略下面的测试.用relion_postprocess计算FSC.用'relion_star_handler --i OUTPUT.STAR --o out.star --do_invert_BBR_handerness'来生成相反手性的star,重构后比较FSC.保留FSC较高的那个.对于C/D对称的样品,建议前面的手性部分选"2",即不改变block defocus.因为一般来说C/D对称的解不到接近Ewald Sphere极限的分辨率.
正常的单颗粒分析流程,最好总是进行局域refine (Initial angular sampling <= Local searches from auto-sampling in 3D auto-refine)
移动老版本的readme.md到https://github.com/homurachan/Block-based-recontruction/blob/master/old_version/Readme.old2.md
在relion_star_handler里更新了--do_rotation功能，使用如下命令："relion_star_handler --i input.star --o output.star --do_rotation --rOt rotangle --tIlt tiltangle --pSi psiangle". 该命令能把input.star所对应的重构按照RELION使用的ZYZ欧拉角体系旋转rotangle-tiltangle-psiangle，适用于relion_align_symmetry的输出结果. 这样就不需要重新跑class3D.