And it’s hard to choose just one, but it would have to be Bonnie Bassler’s talk on bacterial communication, which just went <a href="http://go.ted.com/cY4" rel="nofollow">online</a>.

I had dinner with Bonnie before her talk and was captivated. So much so that I changed seats to talk with her instead of Paul Allen, Bezos, and some other very interesting <a href="https://www.flickr.com/photos/jurvetson/3264312469">folks</a> around the table.

There are 10x as many bacterial cells as human cells in and on your body, and 100x as many bacterial genes as human genes. And you depend on them for survival.

“At best you are 10% human, but closer to 1%, depending on which metrics you like.”

It reminded me of Craig Venter’s observation that the bacteria in each of our lungs is genetically unique, evolving in real time in an adaptive dance with our immune system.

How can some bacteria make us sick? At a fraction of the size of a human cell, “bacteria are too small to have an impact on the environment if they simply act as individuals.”

“All bacteria can talk to each other. They make chemical words, they recognize those words, and they turn on group behaviors that are only successful when all of the cells participate in unison.”

“Bacteria always control pathogenicity with quorum sensing.”

They wait until they are concentrated enough before they launch an attack on their host; otherwise they would be wasting energy when sub-threshold.

And to prevent local hot spots emerging in isolation from their more distant peers, each bacteria has a positive feedback loop, dumping more signaling chemicals into the bath once they cross a threshold (this way, the entire population can switch modes in unison, from growth mode to attack mode for example).

Each bacteria has two separate signaling pathways. One receptor is exquisitely unique to each species, and one general to all bacteria (“bacterial Esperanto”). Bacteria can distinguish self from other and can determine their absolute concentration and their concentration relative to other bacteria.

“This is the invention of multi-cellularity. We think bacteria made the rules for how multi-cellular organization works.”

Jam the receptor. Instead of killing the bacteria, which would engendering evolutionary resistance, “what if we could do ‘behavior modifications’ so they can’t talk and don’t know to launch virulence.”

At dinner, I suggested that artificially induced group behaviors may also be interesting in biofilms used for water purification and other non-medical applications.

I was also reminded of my first course on neural networks (Rummelhardt, PDP). After simulating a number of neuronal circuits, early research concluded that the neuron needs a sigmoid function — some non-linear tipping point in its response curve — for learning to occur. The positive feedback loop in bacterial quorum sensing is just such a sigmoid, and may provide clues to the early development of cellular signaling, initially among nearest neighbors in pre-Cambrian blobs, and then across differentiated body plans with the long-span neuronal cells. I’m not thinking of evolutionary subsumption or endobiosis, but a resonant developmental homology in primitive signaling systems.

“When you learn new things about natural science, whenever you read something miraculous, it was done by a child. Everyone on my research team is between 20 and 30 years old. They are the engine that drives scientific discovery in this country.”