Scripts to run Bill Swope's conformer energy analysis
NC Number of QM conformers
NM Number of QM matches with any of the MM structures
NO Number of QM orphans: QM structures that don’t match with any MM struc. within a threshold
EO 10*(num orph)/(num conf): maximum # of confs is 10, might be an equalization metric for # of orphans
PC Num cases match .ne. p-c: Cases where the QM has a close match to a different MM struc (not a parent-child)
LQ 1 if lowest QM is orphan: If the low energy QM does not have any MM matches
LF 1 of lowest FF is orphan: If the MM minima does not correspond to any QM
BF Num FF strucs below ref: Number of FF structures below the reference FF struc that matches to the QM minima
MO FF strucs out of order: number of FF structures out of order rel to QM order
BE Num small dE large ddE: number of cases where ddE is large but dE is small
QS Num 2Q:1F matches: Cases where more than one QM matches to a single FF struc
FS Num 1Q:2F matches: Cases where more than one FF matches to a single QM struc
SC 0-99 score: (# of QM matches - # of flagged cases with BE)/total_qm_conf, higher the better
NL 0-99 score lowE passes: how many have |ddE|<2 among low energy confs (|dE|<4), higher the better
create_inputs_for_swope_analysis.py - takes in the QM and MM files (both ordered the same, n^{th} molecule in each are the same)
name_map_to_qca_id.json - name map from QCA id to industry benchmark set compound id
computermsd.f - computes the QM x MM RMSD matrix
tfd_matrix_qm_by_mm.py - creates a QM x MM matrix in a format similar to the fortran script which calculates RMSD
computematch.f - computes the fingerprint based on the QM x MM RMSD/TFD matrix for each molecule
computematch.out - compiled executable
computermsd.out - compiled executable
get_mol_with_name.py - utility script to extract a molecule by name
MM-all.sdf - MM optimized conformer geometries in SDF format
qm-all-annotated.sdf - QM reference conformer geometries in SDF format
./test-data - directory created with the create_inputs_for_swope_analysis.py script
Compile the fortran files computermsd.f and computematch.f using gfortran.
gfortran computermsd.f -llapack -o computermsd.out
gfortran computematch.f -llapack -o computematch.out
QM references in one file and corresponding MM reference in the other. This script will create directories in the format needed by the above scripts (compute*)
python create_inputs_for_swope_analysis.py --qm_file_name qm-all-annotated.sdf --mm_file_name MM-all.sdf --iter_name test
Copy the bash script report.sh to the data folder and run it