English subtitles

← The next software revolution: programming biological cells

Get Embed Code
27 Languages

Showing Revision 8 created 11/01/2019 by Oliver Friedman.

  1. The second half of the last century
    was completely defined
  2. by a technological revolution:
  3. the software revolution.
  4. The ability to program electrons
    on a material called silicon
  5. made possible technologies,
    companies and industries
  6. that were at one point
    unimaginable to many of us,
  7. but which have now fundamentally changed
    the way the world works.
  8. The first half of this century, though,
  9. is going to be transformed
    by a new software revolution:
  10. the living software revolution.
  11. And this will be powered by the ability
    to program biochemistry
  12. on a material called biology.
  13. And doing so will enable us to harness
    the properties of biology
  14. to generate new kinds of therapies,
  15. to repair damaged tissue,
  16. to reprogram faulty cells
  17. or even build programmable
    operating systems out of biochemistry.
  18. If we can realize this --
    and we do need to realize it --
  19. its impact will be so enormous
  20. that it will make the first
    software revolution pale in comparison.
  21. And that's because living software
    would transform the entirety of medicine,

  22. agriculture and energy,
  23. and these are sectors that dwarf
    those dominated by IT.
  24. Imagine programmable plants
    that fix nitrogen more effectively
  25. or resist emerging fungal pathogens,
  26. or even programming crops
    to be perennial rather than annual
  27. so you could double
    your crop yields each year.
  28. That would transform agriculture
  29. and how we'll keep our growing
    and global population fed.
  30. Or imagine programmable immunity,
  31. designing and harnessing molecular devices
    that guide your immune system
  32. to detect, eradicate
    or even prevent disease.
  33. This would transform medicine
  34. and how we'll keep our growing
    and aging population healthy.
  35. We already have many of the tools
    that will make living software a reality.

  36. We can precisely edit genes with CRISPR.
  37. We can rewrite the genetic code
    one base at a time.
  38. We can even build functioning
    synthetic circuits out of DNA.
  39. But figuring out how and when
    to wield these tools
  40. is still a process of trial and error.
  41. It needs deep expertise,
    years of specialization.
  42. And experimental protocols
    are difficult to discover
  43. and all too often, difficult to reproduce.
  44. And, you know, we have a tendency
    in biology to focus a lot on the parts,
  45. but we all know that something like flying
    wouldn't be understood
  46. by only studying feathers.
  47. So programming biology is not yet
    as simple as programming your computer.
  48. And then to make matters worse,
  49. living systems largely bear no resemblance
    to the engineered systems
  50. that you and I program every day.
  51. In contrast to engineered systems,
    living systems self-generate,
  52. they self-organize,
  53. they operate at molecular scales.
  54. And these molecular-level interactions
  55. lead generally to robust
    macro-scale output.
  56. They can even self-repair.
  57. Consider, for example,
    the humble household plant,

  58. like that one sat
    on your mantelpiece at home
  59. that you keep forgetting to water.
  60. Every day, despite your neglect,
    that plant has to wake up
  61. and figure out how
    to allocate its resources.
  62. Will it grow, photosynthesize,
    produce seeds, or flower?
  63. And that's a decision that has to be made
    at the level of the whole organism.
  64. But a plant doesn't have a brain
    to figure all of that out.
  65. It has to make do
    with the cells on its leaves.
  66. They have to respond to the environment
  67. and make the decisions
    that affect the whole plant.
  68. So somehow there must be a program
    running inside these cells,
  69. a program that responds
    to input signals and cues
  70. and shapes what that cell will do.
  71. And then those programs must operate
    in a distributed way
  72. across individual cells,
  73. so that they can coordinate
    and that plant can grow and flourish.
  74. If we could understand
    these biological programs,

  75. if we could understand
    biological computation,
  76. it would transform our ability
    to understand how and why
  77. cells do what they do.
  78. Because, if we understood these programs,
  79. we could debug them when things go wrong.
  80. Or we could learn from them how to design
    the kind of synthetic circuits
  81. that truly exploit
    the computational power of biochemistry.
  82. My passion about this idea
    led me to a career in research

  83. at the interface of maths,
    computer science and biology.
  84. And in my work, I focus on the concept
    of biology as computation.
  85. And that means asking
    what do cells compute,
  86. and how can we uncover
    these biological programs?
  87. And I started to ask these questions
    together with some brilliant collaborators
  88. at Microsoft Research
    and the University of Cambridge,
  89. where together we wanted to understand
  90. the biological program
    running inside a unique type of cell:
  91. an embryonic stem cell.
  92. These cells are unique
    because they're totally naïve.
  93. They can become anything they want:
  94. a brain cell, a heart cell,
    a bone cell, a lung cell,
  95. any adult cell type.
  96. This naïvety, it sets them apart,
  97. but it also ignited the imagination
    of the scientific community,
  98. who realized, if we could
    tap into that potential,
  99. we would have a powerful
    tool for medicine.
  100. If we could figure out
    how these cells make the decision
  101. to become one cell type or another,
  102. we might be able to harness them
  103. to generate cells that we need
    to repair diseased or damaged tissue.
  104. But realizing that vision
    is not without its challenges,
  105. not least because these particular cells,
  106. they emerge just six days
    after conception.
  107. And then within a day or so, they're gone.
  108. They have set off down the different paths
  109. that form all the structures
    and organs of your adult body.
  110. But it turns out that cell fates
    are a lot more plastic

  111. than we might have imagined.
  112. About 13 years ago, some scientists
    showed something truly revolutionary.
  113. By inserting just a handful of genes
    into an adult cell,
  114. like one of your skin cells,
  115. you can transform that cell
    back to the naïve state.
  116. And it's a process that's actually
    known as "reprogramming,"
  117. and it allows us to imagine
    a kind of stem cell utopia,
  118. the ability to take a sample
    of a patient's own cells,
  119. transform them back to the naïve state
  120. and use those cells to make
    whatever that patient might need,
  121. whether it's brain cells or heart cells.
  122. But over the last decade or so,

  123. figuring out how to change cell fate,
  124. it's still a process of trial and error.
  125. Even in cases where we've uncovered
    successful experimental protocols,
  126. they're still inefficient,
  127. and we lack a fundamental understanding
    of how and why they work.
  128. If you figured out how to change
    a stem cell into a heart cell,
  129. that hasn't got any way of telling you
    how to change a stem cell
  130. into a brain cell.
  131. So we wanted to understand
    the biological program
  132. running inside an embryonic stem cell,
  133. and understanding the computation
    performed by a living system
  134. starts with asking
    a devastatingly simple question:
  135. What is it that system actually has to do?
  136. Now, computer science actually
    has a set of strategies

  137. for dealing with what it is the software
    and hardware are meant to do.
  138. When you write a program,
    you code a piece of software,
  139. you want that software to run correctly.
  140. You want performance, functionality.
  141. You want to prevent bugs.
  142. They can cost you a lot.
  143. So when a developer writes a program,
  144. they could write down
    a set of specifications.
  145. These are what your program should do.
  146. Maybe it should compare
    the size of two numbers
  147. or order numbers by increasing size.
  148. Technology exists that allows us
    automatically to check
  149. whether our specifications are satisfied,
  150. whether that program
    does what it should do.
  151. And so our idea was that in the same way,
  152. experimental observations,
    things we measure in the lab,
  153. they correspond to specifications
    of what the biological program should do.
  154. So we just needed to figure out a way

  155. to encode this new type of specification.
  156. So let's say you've been busy in the lab
    and you've been measuring your genes
  157. and you've found that if Gene A is active,
  158. then Gene B or Gene C seems to be active.
  159. We can write that observation down
    as a mathematical expression
  160. if we can use the language of logic:
  161. If A, then B or C.
  162. Now, this is a very simple example, OK.
  163. It's just to illustrate the point.
  164. We can encode truly rich expressions
  165. that actually capture the behavior
    of multiple genes or proteins over time
  166. across multiple different experiments.
  167. And so by translating our observations
  168. into mathematical expression in this way,
  169. it becomes possible to test whether
    or not those observations can emerge
  170. from a program of genetic interactions.
  171. And we developed a tool to do just this.

  172. We were able to use this tool
    to encode observations
  173. as mathematical expressions,
  174. and then that tool would allow us
    to uncover the genetic program
  175. that could explain them all.
  176. And we then apply this approach
  177. to uncover the genetic program
    running inside embryonic stem cells
  178. to see if we could understand
    how to induce that naïve state.
  179. And this tool was actually built
  180. on a solver that's deployed
    routinely around the world
  181. for conventional software verification.
  182. So we started with a set
    of nearly 50 different specifications
  183. that we generated from experimental
    observations of embryonic stem cells.
  184. And by encoding these
    observations in this tool,
  185. we were able to uncover
    the first molecular program
  186. that could explain all of them.
  187. Now, that's kind of a feat
    in and of itself, right?

  188. Being able to reconcile
    all of these different observations
  189. is not the kind of thing
    you can do on the back of an envelope,
  190. even if you have a really big envelope.
  191. Because we've got
    this kind of understanding,
  192. we could go one step further.
  193. We could use this program to predict
    what this cell might do
  194. in conditions we hadn't yet tested.
  195. We could probe the program in silico.
  196. And so we did just that:

  197. we generated predictions
    that we tested in the lab,
  198. and we found that this program
    was highly predictive.
  199. It told us how we could
    accelerate progress
  200. back to the naïve state
    quickly and efficiently.
  201. It told us which genes
    to target to do that,
  202. which genes might even
    hinder that process.
  203. We even found the program predicted
    the order in which genes would switch on.
  204. So this approach really allowed us
    to uncover the dynamics
  205. of what the cells are doing.
  206. What we've developed, it's not a method
    that's specific to stem cell biology.

  207. Rather, it allows us to make sense
    of the computation
  208. being carried out by the cell
  209. in the context of genetic interactions.
  210. So really, it's just one building block.
  211. The field urgently needs
    to develop new approaches
  212. to understand biological
    computation more broadly
  213. and at different levels,
  214. from DNA right through
    to the flow of information between cells.
  215. Only this kind of
    transformative understanding
  216. will enable us to harness biology
    in ways that are predictable and reliable.
  217. But to program biology,
    we will also need to develop

  218. the kinds of tools and languages
  219. that allow both experimentalists
    and computational scientists
  220. to design biological function
  221. and have those designs compile down
    to the machine code of the cell,
  222. its biochemistry,
  223. so that we could then
    build those structures.
  224. Now, that's something akin
    to a living software compiler,
  225. and I'm proud to be
    part of a team at Microsoft
  226. that's working to develop one.
  227. Though to say it's a grand challenge
    is kind of an understatement,
  228. but if it's realized,
  229. it would be the final bridge
    between software and wetware.
  230. More broadly, though, programming biology
    is only going to be possible

  231. if we can transform the field
    into being truly interdisciplinary.
  232. It needs us to bridge
    the physical and the life sciences,
  233. and scientists from
    each of these disciplines
  234. need to be able to work together
    with common languages
  235. and to have shared scientific questions.
  236. In the long term, it's worth remembering
    that many of the giant software companies

  237. and the technology
    that you and I work with every day
  238. could hardly have been imagined
  239. at the time we first started
    programming on silicon microchips.
  240. And if we start now to think about
    the potential for technology
  241. enabled by computational biology,
  242. we'll see some of the steps
    that we need to take along the way
  243. to make that a reality.
  244. Now, there is the sobering thought
    that this kind of technology
  245. could be open to misuse.
  246. If we're willing to talk
    about the potential
  247. for programming immune cells,
  248. we should also be thinking
    about the potential of bacteria
  249. engineered to evade them.
  250. There might be people willing to do that.
  251. Now, one reassuring thought in this
  252. is that -- well, less so
    for the scientists --
  253. is that biology is
    a fragile thing to work with.
  254. So programming biology
    is not going to be something
  255. you'll be doing in your garden shed.
  256. But because we're at the outset of this,
  257. we can move forward
    with our eyes wide open.
  258. We can ask the difficult
    questions up front,
  259. we can put in place
    the necessary safeguards
  260. and, as part of that,
    we'll have to think about our ethics.
  261. We'll have to think about putting bounds
    on the implementation
  262. of biological function.
  263. So as part of this, research in bioethics
    will have to be a priority.
  264. It can't be relegated to second place
  265. in the excitement
    of scientific innovation.
  266. But the ultimate prize,
    the ultimate destination on this journey,

  267. would be breakthrough applications
    and breakthrough industries
  268. in areas from agriculture and medicine
    to energy and materials
  269. and even computing itself.
  270. Imagine, one day we could be powering
    the planet sustainably
  271. on the ultimate green energy
  272. if we could mimic something
    that plants figured out millennia ago:
  273. how to harness the sun's energy
    with an efficiency that is unparalleled
  274. by our current solar cells.
  275. If we understood that program
    of quantum interactions
  276. that allow plants to absorb
    sunlight so efficiently,
  277. we might be able to translate that
    into building synthetic DNA circuits
  278. that offer the material
    for better solar cells.
  279. There are teams and scientists working
    on the fundamentals of this right now,
  280. so perhaps if it got the right attention
    and the right investment,
  281. it could be realized in 10 or 15 years.
  282. So we are at the beginning
    of a technological revolution.

  283. Understanding this ancient type
    of biological computation
  284. is the critical first step.
  285. And if we can realize this,
  286. we would enter in the era
    of an operating system
  287. that runs living software.
  288. Thank you very much.

  289. (Applause)