MOLGENGO

Beomchang Kang*, Chaok Seok*, Juyong Lee**

*Seoul National University

**Kangwon National University


User guide

These models are the result of training for prediction of electronic transition as described in Kang et al., "MOLGENGO: Finding Novel Molecules with Desired Electronic Properties by Capitalizing on Their Global Optimization". molgengo.zip contains a binary files. Its name is molgengo. zink_250k.smi and bank_1 are examples of input and output file, respectively. Input file is a list of SMILES formats. In output file, the third column represents the maximum oscillator strength and the second column represents its wavelength. For log message, the third column represents the maximum oscillator strength and the second column represents its wavelength. Finally, os_parameter and ee_parameter are parameter files for LGBM prediction.
usage: ./molgengo
  There are many options
  --smifile [string, smiles file for first bank]
  --seed [integer, ineger for random number generation]
  --mu [integer, bank size]
  --lam [integer, the number of new gens per generation]
  --nu NU [integer, the number of seed genes]
  --generation [integer, the limit of generation]
  --Dsch [float, scheduling parameter]
  --w1 [float, weight for wavelength]
  --time [integer, the limit of time(hour)]

Download

References

  1. Kang, Beomchang; Seok, Chaok; Juyong Lee. MOLGENGO: Finding Novel Molecules with Desired Electronic Properties by Capitalizing on Their Global Optimization, ACS Omega 2021 6 (41), 27454-27465.

Contact

[email protected]