There are a couple of pretty big hurdles I'm going to have to cross pretty soon. This includes, but is not limited to,
An accurate way of determining the pKa of a molecule (and/or its conjugate acid/base)
An accurate way of determining what, if any, resonance structures a molecule has
An accurate way of determining whether or not a molecule is aromatic or anti-aromatic, and where it exhibits these trends
These are all challenging tasks. I don't know that I'm up to them, and in general I don't know how well software can predict some of these things in general (I'm certain that they're doable to an extent, but it probably requires a greater understanding of both computer science and chemistry than I have).
Thus, I seem to be left with only one option. Find existing databases and information on molecules (such as ChEMBL) and use them as the first layer. Then if information regarding a molecule cannot be found within one of those, fallback to my (likely less than stellar) algorithmic implementation of them. I can also generate my own database and add items to it as necessary (although I hope it won't be, at least because "[ChEMBL] currently contains over 1.4 million unique structures with the associated activity at 10,579 different targets".
I'm going to continue searching for open source packages that I can use to help with the various issues I've been coming across. If/when I can get openbabel working on windows then I'll start with improved file parsing, and then move onto using openbabel's data structures and functions to replace many of mine.
There are a couple of pretty big hurdles I'm going to have to cross pretty soon. This includes, but is not limited to,
These are all challenging tasks. I don't know that I'm up to them, and in general I don't know how well software can predict some of these things in general (I'm certain that they're doable to an extent, but it probably requires a greater understanding of both computer science and chemistry than I have).
Thus, I seem to be left with only one option. Find existing databases and information on molecules (such as ChEMBL) and use them as the first layer. Then if information regarding a molecule cannot be found within one of those, fallback to my (likely less than stellar) algorithmic implementation of them. I can also generate my own database and add items to it as necessary (although I hope it won't be, at least because "[ChEMBL] currently contains over 1.4 million unique structures with the associated activity at 10,579 different targets".
I'm going to continue searching for open source packages that I can use to help with the various issues I've been coming across. If/when I can get openbabel working on windows then I'll start with improved file parsing, and then move onto using openbabel's data structures and functions to replace many of mine.