The Network Coding and Reliable Communications Group, Professor Muriel Medard

Events

MIT-Harvard Communications Information Networks Circuits and Signals (CINCS) / Hamilton Institute Seminar

Wednesday, October 21, 2020, 11AM, Zoom

Speaker: Damien Woods

Title: 21 molecular algorithms using reprogrammable DNA self-assembly

Abstract: The history of computing tells us that computers can be made of almost anything: silicon, gears and levers, neurons, flowing water, interacting particles or even light.  Although lithographically patterned silicon surfaces have been by far the most successful of these, they give us a limited view of what computation is capable of. Algorithms typically control the flow of information by controlling the flow of electrons through digital-electronic devices, but in the field of molecular computing we imagine algorithms that control matter itself.

This talk will be about DNA molecules that interact with each other in a test tube to execute algorithms.  We will show how DNA can be re-engineered to act not only as an information encoding polymer, as seen in biology, but also as a computational primitive. The talk will showcase our wet-lab results on implementing 21 different algorithms using self-assembling DNA strands [1].  The DNA strands stick together to form a lattice structure, all while executing a so-called 6-bit iterated Boolean circuit.  Our set of 355 DNA strands is reprogrammable: it can be used to implement any 6-bit iterated Boolean circuit.  The per-step mean error rate is 1/3000, which we contend is low for engineered self-assembling molecules.  Finally, we use ideas from the theory of computation to help us understand what kinds of computations such self-assembling molecules are capable of in general.

[1] Woods*, Doty*, Myhrvold, Hui, Zhou, Yin, Winfree. Diverse and robust molecular algorithms using reprogrammable DNA self-assembly.  Nature 567:366-372.  2019.  *Joint lead co-authors.

Bio: Damien Woods is at the Hamilton Institute, Maynooth University, Ireland, where his group conducts research on molecular computers: collections of carefully engineered DNA molecules that bump into each other and interact in a test tube to solve some mathematically specified computational problem. His focus is on both the underlying computational theory and implementation in the lab. The work is funded by the European Research Council (ERC) and Science foundation Ireland (SFI).

He has traveled for a bit, doing research in Inria (France, 2016-2018), Caltech (USA, 2009-2016), University of Seville (Spain) and University College Cork (Ireland).  His PhD is from Maynooth University (Ireland).

He enjoys whiteboarding, pipetting, AFMing and other scientific sports.

Please email mkruko@mit.edu for the zoom link.