Here we keep all the molecular structures we use for the ModelSEED
Biochemistry Database. We download and use molecular structures from
KEGG, MetaCyc and MetaNetX, each stored as SMILES, InChI, and InChIKey
representations in their own folder, and within the
All_ModelSEED_Structures.txt file.
We use Marvin to protonate each mol file at a pH of 7 cannot handle every single file for various reasons, so all files that contain the word "Original" were generated from the set of uncharged molecular structure files that were downloaded from their respective sources, and all files with the word "Charged" concern the files that were sucessfully protonated by Marvin.
Marvin was also used to generate the pKas and pKbs of the atoms in each molecular structure.
We also parsed the KEGG molecular structure files for SRUs (structural repeating units). The information stored for these are passed over by Marvin when protonating, and not repeated in the resulting output structure. This caused problems when merging compounds that wre clearly not the same compound, so we keep a record of which of the KEGG molecular structure files contain these SRUs.