Skip to content

Latest commit

 

History

History
115 lines (77 loc) · 4.63 KB

analysis.md

File metadata and controls

115 lines (77 loc) · 4.63 KB

/* vim: set filetype=markdown : */

Analyzing Quanformer Results

Document last updated: Nov 30 2018
This file describes and guides the user through various analyses that can be conducted from different implementations of Quanformer.

Comparing sets of geometry optimizations

match_minima.py prerequisites

  • SDF files for all methods, where each SDF file has all mols and confs for one particular method
  • input text file listing these SDF files and associated level of theory

match_minima.py output

  • .dat file with summarized RMS errors by level of theory (one dat file per reference method)
  • bar plot with conformer averaged wall times for all methods (one plot per molecule)
  • line plot with conformer relative energies for all methods (one plot per molecule)

match_plot.py prerequisites

  • .dat files from match_minima.py with summarized RMS errors by level of theory (one dat file per reference method)
  • input text file listing these .dat files and associated reference level of theory
  • (optional) input file can be a single .dat file if only wanting a single scatter plot

match_plot.py output

  • heat map of root mean square errors of all methods compared to each other (one plot)
  • heat map of relative times (log ratios) of all methods compared to each other (one plot)
  • scatter plot of log ratio of relative times vs rmse (one plot per reference method)

Instructions See README.md file in examples/match_minima/ in this repo.


Comparing sets of single point energy calculations

The relative energies for a set of molecules/conformers can be evaluated among various QM methods. The same starting SDF file should be used in the various QM calculations, and no conformers should be "lost" (such as from redundant structures, incomplete optimizations). This section is closely related to the one above of comparing sets of geometry optimizations, but it does not need to ensure matching structures.

Prerequisites

  • TODO

Instructions

  1. Create input file for stitchSpe.py -- see notes below.
  2. python stitchSpe.py -i /path/and/input.dat --plotbars [--reffile filename.sdf]
  3. (opt.) If some mol has a high RMSD, identify the outlying conformer and visualize structure. (see writeOneMol in examples)

Output

  1. Data file with relative energies, either by RMSD (if there is a reference file) or with avgs/stdevs (if no reference file)
  2. barplot.png

Creating input file for analysis with stitchSpe.py

  • This should be a text file directing the script to process a particular quantity.
  • The SDF files on each line should ALL have the same molecules, same conformers, etc. These may differ in coordinates or SD tags.
  • If you have one file that should be taken as reference for RMSD calculations of energy, specify that in the python call with --reffile [filename].
  • Each line should contain these columns, comma-separated:
    1. Name of SDF file with full path
    2. Calculation type, either: 'spe' or 'opt'
    3. QM method
    4. QM basis set
  • Example input file:
 # comments begin with pound symbol and are ignored
 /path/and/setofMols-221-opt2.sdf, opt, b3lyp-d3mbj, def2-tzvp
 /path/and/setofMols-221-spe1.sdf, spe, b3lyp-d3mbj, def2-tzvp
 /path/and/setofMols-221-spe2.sdf, spe, mp2        , cc-pvtz
 /path/and/setofMols-221-spe3.sdf, spe, pbe0       , 6-311g**

Comparing energies before and after a set of geometry optimizations

Prerequisites

  1. opt_vs_spe.py -- [TODO write instructions]

Instructions

  • TODO

Interfacing with modified Seminario method

The modified Seminario method is a way to obtain force field parameters (force constants and equilibrium values) based on the connectivity and Hessian of a molecule.

Prerequisites

  • SDF file with structures of all Hessian calculations
  • Pickle file with extracted Hessian data (from calling results with executor.py)

Instructions
Follow the three steps of preparing the modified Seminario code detailed at the top of quan2modsem.py.
Then call the code. Example:

  • python quan2modsem.py -i carbon-hess.sdf -p carbon-hess.hess.pickle

Output
Each conformer's directory will have the three files:

  • Modified_Seminario_Angle
  • Modified_Seminario_Bonds
  • MSM_log

References