====== Pre-processing of dataset from Titan II ======

==== 1. Recover the dataset from the microscope ====

Log into **cbi-gateway-01** to access the microscope and transfer the data to **gateway** or **storage**

<code>
ssh <login>@cbi-gateway-01
ls /mnt/titan2/offloaddata/TemScripting/EF-Falcon/<dataset>
rsync -avz --progress /mnt/titan2/offloaddata/TemScripting/EF-Falcon/<dataset> /mnt/storage/teams/<your_team_folder>/<project>/<dataset>
</code>

Path to gateway as seen from cbi-gateway-01 : <quote>/mnt/zfspool/teams/</quote>

==== 2. Convert the EER movies to TIF ====

Log into **cbi-compute-01** and transfer the scripts {{ ::eer2tif_parallel_loop.py |}} and {{ ::eer2tif_parallel.py |}} in your **home**

<code>
ssh <login>@cbi-compute-01
module load relion
cd /mnt/storage/teams/<your_team_folder>/<project>/<dataset>/ #go to the folder with the dataset
module load python
python ~/eer2tif_parallel.py
</code>

Warning : conversion of eer to tiff file will show error this message : <quote>TIFFReadDirectory: Warning, Unknown field with tag 65002 (0xfdea) encountered.</quote>

==== 3. Begin processing with relion ====

Load and start relion (from cbi-compute-01 or the team GPU node (phantom-node39 for Lamour-Ruff))
<code>
module load relion
relion& #start relion in the folder with movies
</code>

=== Job Relion/Import ===
Give the path to the *.tif , specify the pixel size, voltage and hit **RUN**
You can **RESUME** the job when more movies are transferred and converted

=== Estimation the camera gain ===

Estimate the gain with at least 200 movies (redo if needed with a larger number)
<code>relion_estimate_gain --i Import/job001/movies.star --j 8 --max_frames 10000 --random true --o estimated_gain.mrc</code>
Explanation of options:
  * --j: number of thread
  * --max_frames: target number of frames to average (rounded to movies)
  * --random: randomize the order of input movies before taking subset

=== Job Relion/Motion correction ===
 
Specify the dose per frame, the number of patches 5x5m, provide the gain file, and set several MPI (number of tasks in parallel) and several (as a multiple of the number of frames in the movies) threads (number of cpu per task)