For the last few days I have been trying to figure out how I can do a simple Monte Carlo simulation on my computer without much success.
In the last few months I have become convinced that by doing my own computer modeling I could probably avoid making a lot of unnecessary molecules. Indeed, I see no reason why the notion hasn’t struck me in the same way the concept of checking a sample after a reaction for purity why should I check the reaction before I do it to see how well it might work?
Firstly, it’s become pretty obvious that computational chemistry is a giant fucking black box. There are lots of force fields and data sets and letters followed by ** and shit and not a goddamn easy way to deal with ANY of it. Let’s say I want to model a transition state, what do I do? Call a computational chemist? I’m sure they’ve got shit they want to do, too. Not only that, but after a couple decades of ripening, I’d like to think the ability to do simple computational analysis should be in the hands of the average schmuck like me.
So, I want to do a Monte Carlo simulation. This, as I have come to appreciate it, simply means I want to find a global minimum for a structure – i.e. the lowest energy configuration something sits in, but there’s no easy way to figure out how to do this. In so far as I can tell, you can do the following things:
Draw something in chemdraw and either import it into Chem3D if you own a PC or into Avogadro if you own a Mac (which will convert chemdraw .cdx files into 3D coordinate files). Do a minimization using whatever they have. Avogadro is clearly better at this (and it works for PC) and has a number of different choices. The problem with both of those program is that they really don’t find a global minimum. Insofar as I can tell, if you start with a shitty structure, both of them will find a minimum of that shitty structure. True Monte Carlo, as I understand it, means it will rebuild the structure and keep searching.
Then I discovered a program called Tinker… it has a bunch of force fields and shit in it, all prime and ready to use (and it even works on Windows and Mac) but no clear path from Avogadro or Chem3D to a minimization is presented.
I don’t understand why computational chemists don’t want other chemists to do computational chemistry with a graphical user interface and dialog boxes and all those smart things that can be made using Visual Basic or Apple script… It just seems to me that, a lab like Jay Ponder‘s or Dan Gezelter’s could really make a difference by making a user application that would allow the simple importation of a Cartesian coordinate file and allow the user to select a Force Field, Algorithm and a good stopping point to do their own “back of the napkin” calculations before they run off into the lab… If I had the wherewithal to do it, I’d have done it by now.
UPDATE: Re: Collaborate with a computational chemist:
I have done that before, put his name on the paper and everything. The problem is that I’m *JUST* starting out in my little Post-Doc adventure and so I don’t really have systems in mind that I would want to bother someone with. I.E. some of them are likely patently absurd – simply won’t work for reasons I don’t need a computer to tell me… but I have a chemdraw folder full of them and a billion other (more pressing) things to be doing. I also have a pretty ass kickity computer that has no problem crunching numbers all day, both at home and at school.
I also want to obtain some degree of self sufficiency. I feel as though, in this modern era, I don’t need to run to an NMR chemist to take a sample of, what will normaly turn out to be ethyl acetate, so why should I need to run to a computational chemist to do something so very trivial. It’s a waste of her time to do my calculations for me on a compound that has a 1% shot of even being made!