SoylentNews is people
SoylentNews
This high-throughput reaction prediction (HTRP) idea has been tried several times before, and the paper provides a useful review of these. Broadly, these attempts have used either some sort of rule-based expert system framework, attempted to work out a logic or grammar of chemical reactivity to extrapolate with, or used outright machine learning techniques.
[...] As a number of other theoretical approaches to organic chemistry have done, they're shifting the world of organic chemistry over into a graph-theory problem. From this perspective, discovering new reactions becomes a search for new nodes and edges in the graph[...] let's say you have two molecules, A and B, and both of them are both known to react with a partner C to give some new product in each case. The program will note all these similarities, and searches for cases where compound A reacts with yet another molecule type D. Since A and B have been classed as having a similar reactivity pattern (they both reacted usefully with C), the program hypothesizes that B will do something with D as well
[...] They tested this approach by using everything up to 2013 to predict reactions, and the set of reactions published since then to check their results. Looking at 180,000 randomly selected reactions, the predictions were correct about 67% of the time.
http://blogs.sciencemag.org/pipeline/archives/2016/08/31/predicting-new-reactions
http://arxiv.org/abs/1608.07117
(Score: 3, Interesting) by quietus on Sunday September 04 2016, @10:42AM
In 2013, a train carrying 300 ton of acrylonitril and 1,3-butadiene derailed near Wetteren, Belgium.
6 wagons got off the tracks, 3 started burning; releasing fumes of acrylonitril and pyrolysis by-products (hydrogen cyanide, nitrogen oxides and ethyn (acetylene)). Fire services arrived, and managed to extinguish the chemical fire. To prevent the other wagons from exploding, they cooled them with water. Part of the acrylonitril had dissolved in the cooling water, which ended up into the municipal sewage system of the nearby village, where oxidation and hydrolysis created cyanide-containing gases. These invaded people's houses through the toilet(s): one person died, and the complete nearby village had to be evacuated -- with the last inhabitants only able to return to their homes 3 weeks later.
Part of the problem was that the fire services didn't know exactly what they were trying to extinguish for nearly 2 hours (fire services inside information). This is not an uncommon problem with chemical fires. Which brings us to problem number two: there are currently no sensors on the market -- as far as I know -- which can detect more than a couple of classes of organic or chemical compounds.
The research/computational chemistry approach discussed here is interesting in this respect: you could keep the graph approach, and combine it with a limit in the number of nodes involved (fire extinguishing chemicals used by the fire services, most common transportation forms of organic chemicals used in the plastics industry, and so on), to give at least an indication of the risks involved with each approach in fire fighting.