Documents and data for EMPIAR workshop

Analyzing 2D image data stored in EMPIAR often requires large computational costs and detailed knowlege about single pariticle analysis. For beginners for single particle analysis, we open the tutorial documents and data used for workshops organized by PDBj.

Innexin-6 from EMPIAR-10291

The images stored as EMPIAR-10291 collected on JEOL JEM-3000SFF(G3), provides a high resolution (3.6 angstrom) 3D density map for the membrane protein innexin-6. This data set contains 300 motion-correced still micrographs (15.9 GB). To reduce computation costs, we prepared 100 micrographs and 50 micrographs data set. Both datasets provides a map with a reasonablly high resolution (3.96 angstrom). We thank an original author,Prof. Atsunori Oshima to advice us to analyze their data.

EMPIAR-10291 EMPIAR-10291_100mic EMPIAR-10291_50mic
Filesize 15.9 GB 5.4 GB 2.7 GB
Images 300 micrographs 100 micrographs 50 micrographs
3Dmap EMD-9973
Resolution 3.6 angstrom 3.96 angstrom 3.96 angstrom
The original EMPIAR-10291 (15.9 GB) contains 300 mrc files, each of them has 3710 x 3838 x 1 (#MRC_MODE 2 "32-bit signed real').
  1. Execute "CTFfind" by Relion 3.1 (CTFFIND-4.1).
  2. Check CtfFind/*/micrographs_ctf.star First, we checked a distribution of _rlnCtfFigureOfMerit #8 and _rlnCtfMaxResolution #9, and cofirmed it has a single-peak distribution. Then, we decided to choose good micrographs only by MaxResolution.
We have organized two workshops using this tutorial data:

Apo-ferritin from EMPIAR-10248

The images stored as EMPIAR-10248 collected on JEOL CRYO ARM 300, provides a very high resolution (1.54 angstrom) 3D density map for apo ferritin. However, this data is not good for educational workshops and exercises because analyzing this data set (145.9 Gbyte) requires large computational costs. We therefore provide a reduced data set (1.7 Gbyte) for an educational purpose.

EMPIAR-10248 EMPIAR-10148_tutorial
Filesize145.9 GB1.7GB
Movies50 frames x 971 files25 frames x 20 files
3D map EMD-9865
Resolution1.54 angstrom2.84 angstrom
Computational timeMore than 1 week using 4GPUs1.5 hour using 2GPUs
Procedure to reduce image data
  1. For the 971 TIFF files in EMPIAR-10248, we executed MotionCorr and CTFfind using Realion 3.0.
  2. Among the TIFF fiels of EMPIAR-10248 with _rlnCtfFigureOfMerit > 0.1, the 20 TIFF files with the best _rlnCtfMaxResolution have been taken.
  3. Each TIFF file has 50 frames. We took only first 25 frames.