Skip to content

Latest commit

 

History

History
329 lines (231 loc) · 9.54 KB

README.md

File metadata and controls

329 lines (231 loc) · 9.54 KB
Raw

ST-on-the-fly

Simulated Tempering with fast on-the-fly weight determination. Inspired from Nguyen 2013.

Disclaimer

This work is still ongoing, and seems to not work properly at the moment.

Overview

The goal of this project is to provide a tool to run Simulated Tempering. It implements algorithm described in N'Guyen 2013.

Contributor :

This work was initiated for a short Python project given in 2nd year of master deggree of bioinformatics. Please contact Athenais VAGINAY at [email protected] for all inquiries.

Requirements

TODO : check if the code is functional with Python3

Installation

Just download the archive in and untargz it :

cd <path> 
tar zxvf VAGINAY.athenais.tar.gz

Testing

Each module can be tested (doctest) running the module as a script :

cd code
python <module_name>

This causes the examples in the docsttrings to get executed and verified.

TODO : finish unit test

Code info:

The code is not fully PEP8 compliant.

Documentation :

You can use pydoc outputed documentation. html files are store in ./code/doc

Usage

/!
Not not cheat on arguments because combinations are not verified in the code.

TODO : verify arguments combination

The script should be able to recognize abbreviation of long option.

Run Simulated Tempering (ST)

cd code
python ST-on-the-fly.py [-h] --Tmin TMIN --Tmax TMAX --Tnum TNUM
                        [--gro-filename GRO_FILENAME] [--top-filename
                        TOP_FILENAME] --nb-run NB_RUN --simu-type {md,mc}
                        [--st-mdp-template-filename ST_MDP_TEMPLATE_FILENAME]
                        --st-outname ST_OUTNAME
                        [--minimisation | --no-minimisation]
                        [--minimisation-mdp-filename MINIMISATION_MDP_FILENAME]
                        [--minimisation-outname MINIMISATION_OUTNAME]
                        [--gene-veloc-outname GENE_VELOC_OUTNAME]
                        --out-path OUT_PATH [--maxwarn MAXWARN]
                        [--clean-all] [-v[vv]]

Arguments Explanation :

  • -h, --help show this help message and exit
  • --Tmin TMIN At which temperature (in Kelvin) begins the experiment
  • --Tmax TMAX The maximal temperature (in Kelvin) during the experiment
  • --Tnum TNUM Number of temperature in the given range [Tmin, Tmax]
  • --gro-filename GRO_FILENAME : gro file to use for the ST experiment
  • --top-filename TOP_FILENAME : topology file to use for the ST experiment
  • --nb-run NB_RUN The number of molecular dynamics during the experiment :t_one-md * num-run = t_ST
  • --simu-type {md,mc} type of simulation : md - Molecular Dynamics (currently the only working choice) mc - Monte Carlo (currently not working)
  • --st-mdp-template-filename ST_MDP_TEMPLATE_FILENAME mdp file to use for the ST experiment
  • --st-outname ST_OUTNAME template name for output of the ST experiment Should not match either minimisation-outname nor gene-veloc-outname
  • --minimisation Run a minimisation before ST experiment
  • --no-minimisation Do run a minimisation before ST experiment
  • --minimisation-mdp-filename MINIMISATION_MDP_FILENAME mdp file to use for minimisation _ IF MINIMISATION _
  • --minimisation-outname MINIMISATION_OUTNAME template name for output of minimisation Should not match either st-outname nor gene-veloc-outname _ IF MINIMISATION _
  • --gene-veloc-outname GENE_VELOC_OUTNAME template name for output of velocities generation Should not match either st-outname nor minimisation-outname _ IF MINIMISATION _
  • --out-path OUT_PATH Where the outputed results files should be store /!\ must finish by '/'
  • --maxwarn MAXWARN The max number of warnings allowed when running MD
  • --clean-all Clean gromacs outputs unecessary to plot the figures in N'Guyen 2013 /!\ Be carefull : NOT WORKING
  • -v, --verbose Turn on detailed info log The level of verbosity is function of the v count

using Molecular Dynamics with GROMACS

without minimisation

cd code
python ST-on-the-fly.py [-h] --Tmin TMIN --Tmax TMAX --Tnum TNUM
                        --gro-filename GRO_FILENAME --top-filename
                        TOP_FILENAME --nb-run NB_RUN --simu-type md
                        --st-mdp-template-filename ST_MDP_TEMPLATE_FILENAME
                        --st-outname ST_OUTNAME
                        [--no-minimisation]
                        --out-path OUT_PATH [--maxwarn MAXWARN]
                        [--clean-all] [-v[vv]]

Working example :

cd code
python ST-on-the-fly.py \
--Tmin 1 --Tmax 5 --Tnum 1 \
--gro-filename ../data/ala10_md000.gro \
--top-filename ../data/ala10.top \
--nb-run 3  --simu-type md \
--st-mdp-template-filename ../data/md1.mdp  --st-outname outst  \
--out-path ../result/test-without-minimi/   --maxwarn 1       -vvv

with minimisation

cd code
python ST-on-the-fly.py [-h] --Tmin TMIN --Tmax TMAX --Tnum TNUM
                        --gro-filename GRO_FILENAME --top-filename
                        TOP_FILENAME --nb-run NB_RUN --simu-type md
                        --st-mdp-template-filename ST_MDP_TEMPLATE_FILENAME
                        --st-outname ST_OUTNAME
                        --minimisation
                        --minimisation-mdp-filename MINIMISATION_MDP_FILENAME
                        --minimisation-outname MINIMISATION_OUTNAME
                        --gene-veloc-outname GENE_VELOC_OUTNAME
                        --out-path OUT_PATH [--maxwarn MAXWARN]
                        [--clean-all] [-v[vv]]

Working example :

cd code
python ST-on-the-fly.py \
--Tmin 1 --Tmax 5 --Tnum 1 \
--gro-filename ../data/ala10_md000.gro \
--top-filename ../data/ala10.top \
--nb-run 3  --simu-type md \
--st-mdp-template-filename ../data/md1.mdp  --st-outname outst  \
--minimisation \
--minimisation-mdp-filename ../data/mini2.mdp  --minimisation-outname miniout \ 
--gene-veloc-outname gene-vel \
--out-path ../result/test-with-minimi/  --maxwarn 1  -vvv

using Monte Carlo

Not working

TODO

Output visualisation

Use :

python plot.py ...

TODO

File formats

Input :

Classical GROMACS files

  • gro file
  • top file
  • mdp file

You can find a description of the format here

Output :

Output of code/md.py

Classical GROMACS files

  • cpt file
  • edr file
  • gro file
  • log file
  • mdp file
  • top file
  • tpr file
  • xtc file

You can find a description of the format here

Output of code/st.py

Classical GROMACS files

  • cpt file
  • edr file
  • gro file
  • log file
  • mdp file
  • top file
  • tpr file
  • xtc file
  • xvg file

You can find a description of the format here

Results file :

  • /.results

tab separated file with following fields :

  • idx : int < - attempt index
  • t_current : float - time at the begining of the run
  • T_current : float - temperature of the run
  • E_run : float - energy of the run
  • E_T : float - updated energy for T_current
  • For each temperature in provided T range : f(T)

Charts :

TODO : charts

Known Bugs :

  • Probabilities of attempt given by the Metropolis Criteriion is always 0 or 1 could be because energies are to big

  • The code is vulnerable to TOCTTOU (Time Of Check to Time Of Use) Error

References

Communication: Simulated tempering with fast on-the-fly weight determination. Nguyen PH1, Okamoto Y, Derreumaux P. 2013