The reason I have my current job is related to water clusters (I'm a SWE at a large internet firm). About 15 years ago I read about pagerank and graph centrality while doing molecular dynamics simulations of DNA and RNA in water. I noticed in my simulations that water would "cluster" at specific locations around the DNA, but I could only really see density, not the "structure". I realized you could create "graphs" of water clusters from the MD trajectories and then analyze them from the perspective of graph centrality. Unfortunately, nobody in academia/government labs had the infrastructure to do this kind of analysis (it doesn't fall under the normal HPC or supercomputer in academia) so I joined a large internet company to get access to their technology and computers. I ended up hiring some scientists to run MD simulations on spare cycles (sort of like Folding @ Home, but inside the datacenter) and we did some cool analysis (although never really did pagerank analysis of water clusters... I did hear somebody external did that, although I don't think it went anywhere).