TRPC4 tutorial data set

To demonstrate how TranSPHIRE works, a TRPC4 tutorial test data set can be downloaded here: https://ftp.gwdg.de/pub/misc/sphire/TranSPHIRE/TranSPHIRE_1.5_trpc4_tutorial.zip

Unzip the downloaded file TranSPHIRE_1.5_trpc4_tutorial.zip:

>>> tar -xf TranSPHIRE_nature_communications.zip

This extracts a folder called TranSPHIRE_1.5_trpc4_tutorial.zip.

The following instructions can also be found in the included README.rst.

Operating systems

The software has been tested on the following operating systems:

Ubuntu 18 Ubuntu 20 CentOS 7

Software dependencies

TranSPHIRE is dependent on different programs:

Included in this zip file

|TranSPHIRE| version 1.5.0
SPHIRE version 1.3_transphire
SPHIRE GPU ISAC version 1.1

Required non-standard hardware

  • Nvidia GPU

Installation

The installation of TranSPHIRE takes about 5 to 20 minutes, depending on the download speed.

Install SPHIRE_v1.3_transphire

>>> bash ./install_sphire_v1.3_transphire.sh

Install GPU ISAC

Make sure that you have CUDA available then run:

>>> bash ./install_gpu_isac.sh

Create a new conda environment for TranSPHIRE

>>> bash ./install_transphire.sh

Put the TranSPHIRE installation in your PATH

>>> export PATH=$(realpath sphire_v1.3_transphire/envs/transphire/bin):${PATH}

Install other dependencies

Please install the other dependencies from the Software dependencies section.

Demo

A TRPC4 Demo data set containig 120 micrograph movies is coming within the ZIP file.

The expected output can be found in the TRPC4_demo_results_expected folder.

The expected runtime of the demo data is:

  • 4.5 hours without 3D refinement on 6 cores.
  • 5.5 hours with 3D refinement on 6 cores.

On a “normal” GPU machine:

  • Intel(R) Xeon(R) CPU E5-2643 v4 @ 3.40Ghz
  • 6 cores / hyperthreading 12 cores
  • 128 GB RAM
  • 2x GeForce RTX 2080 Ti

Instructions

A more detailed version of the instructions is currently in preparation at transphire.readthedocs.io.

  1. Open the TranSPHIRE GUI
>>> transphire --root_directory $(realpath .)
  1. Click the Settings tab.
  2. Click the Input tab.
  • Input project path for frames: click the folder icon and choose the TRPC4_demo folder
  • Input project path for jpg: click the folder icon and choose the TRPC4_demo folder
  • Input frames extension: tiff
  • number of frames: 50
  1. Click the Output tab.
  • Project name: TRPC4_demo_results
  • Rename prefix: TRPC4_
  • Rename suffix: _demo
  1. Click the Global tab.
  • Pixel size: 0.85
  • Cs: 0.001
  • Gain: click the folder icon and choose the TRPC4_demo/gain_ref.mrc file
  • Protein radius: 120
  1. Click the Copy tab.
  • Delete data after Import: Symlink

  1. Click the Path tab.

    Provide the file path values from your downloaded dependencies.

  2. Click the Motion tab.

  • -FmDose: 1.77
  • -Patch: 5 5 20
  1. Click the CTF tab.
  • –f_start: 40
  • –f_stop: 4
  1. Click the Picking tab.
  • –conf*: click the folder icon and choose the TRPC4_demo/config_2020_07.json file
  • –weights*: click the folder icon and choose the TRPC4_demo/gmodel_phosnet_202005_N63_c17.h5 file
  1. Click the Class2d tab and go to the Advanced tab.
  • Nr. Particles: 5000
  • –img_per_grp: 50
  • –minimum_grp_size: 30
  • MPI processes: Choose the number of your physical cores available.
  1. Click the Select2d tab.
  • –weights: click the folder icon and choose the TRPC4_demo/config_2020_07.json file
  1. Click the Auto3d tab.
  • –mpi_procs: Choose the number of your physical cores available.
  • –mpi_submission_command: bash
  • –mpi_submission_template: click the folder icon and choose the submission_bash_template.sh file
  • –memory_per_node: Adjust depending on your system
  • –mol_mass: 900
  • –symmetry: c4
  • input_volume: click the folder icon and choose the TRPC4_reference.hdf file
  • Use SSH: False

Advanced

  • Minimum classes: 0
  • Minimum particles: 0
  1. Click the Start button.