EPIGENETICS AND ANTIBIOTICS
“Hidden Switches in the Mind,” by Eric J. Nestler, discusses epigenetic changes—alterations to how genes behave that do not affect the information they contain. Is it possible that such changes are at least partially responsible for bacteria becoming resistant to various drugs, given that the changes are passed on to daughter cells? If so, the changes would provide yet another way to overcome resistance to various drugs. Instead of looking for an entirely new antibiotic, it might be simpler to find a way to undo the epigenetic changes and restore the bacterial susceptibility to the drugs we already have.
Berkeley Heights, N.J.
Editors’ note: The author, not being a microbiologist, referred this question to Richard Losick, whose laboratory at Harvard University focuses on bacteria. Losick’s reply follows:
Epigenetics does indeed contribute to antibiotic resistance in bacteria by giving rise to bacteria known as persisters. Indeed, epigenetic mechanisms were initially discovered in bacteria, although the mechanisms are quite different from the histone-based ones described in the article (bacteria do not have histones). Persisters are bacteria that survive antibiotic treatment without having acquired a resistance mutation. Instead they have reversibly entered a state in which they are less susceptible to killing by the antibiotic than other genetically identical cells in the population. Indeed, if we could devise drugs that blocked entry into the persister state, such drugs could contribute to the effectiveness of antibiotic therapy.
NOT OURS TO SEE?
Whereas David Weinberger’s speculations about predictive abilities of big data–crunching models in “The Machine That Would Predict the Future” are intriguing, planners and social scientists aren’t about to step aside just yet. As an example of “big data,” IBM’s Watson has impressive computing power when the question is clear, but important societal questions rarely are. For the near future, we don’t see large computing power successfully responding to the simple questions facing modern societies with complex answers: For instance, how do you motivate Asian governments to take action on climate change? How do you reduce poverty? How do you get people out of their cars and onto public transit?
The challenge to prediction today is successfully integrating philosophy and the social and behavioral sciences with the physical sciences and engineering. Just witness the failure of climate scientists to advance the climate change agenda, resulting, in part, from social, behavioral and political scientists being left out of the conversation. With multidisciplinary cooperation as a starter, “big data” might be better equipped to predict the future.
David R. Hardy
I would suggest that there is an insurmountable hurdle to physicist Dirk Helbing’s work, described by Weinberger, in trying to make a “computing system that would effectively serve as the world’s crystal ball”: the discrete architecture of the natural world. Helbing’s background is apparently the modeling of highway traffic, which has a basic linear architecture. Road traffic acts like a hydraulic problem, where small particles can flow into one another continuously. My background is railroads, which couldn’t behave more differently. On almost every level, their options and costs are effectively discrete. Railway costs are highly correlated, irregular, stepwise functions. They are dynamically unstable as they interact. That is, these costs are complexly unique lookup tables, not continuous equations, which means that highway and other linear models cannot be used rigorously (although people do try to use them).