yccai/fragment_mapper

Utilities for finding the smallest set of fragments a molecule is made of using RDKit

Trys to find the smallest set of fragments necessary to cover a
molecule using a greedy algorithm.

Prereqs:

RDKit : github.com/rdkit/rdkit

Run like this:

python is_feasible_mol.py -f fragments.csv molecules.csv mapped_molecules.csv

The mappings are stored in mapped_molecules.csv.

The following PDF reports are generated as well:

Success.pdf - Molecules that could be successfully mapped and the
fragments they match against

MappingNotFound.pdf - Molecules that are completely covered by the
fragments, but a non-overlapping mapping could not be found (for
example, two rings would be required to be fused together)

NotFullyCovered.pdf - Molecules that have atom topologies completely
missing from the fragment set, the missing portion is highlighted in
red.

TimeLimitExceeded.pdf - Molecules that exceeded the 10 second search
limit (haven't seen this occur recently since major tweaks and
optimizations to the algorithm).