BEGIN:VCALENDAR
VERSION:2.0
X-WR-CALNAME:other Events
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080707T140000
DTEND;TZID=US/Eastern:20080707T153000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1910
SUMMARY:Separations between Nondeterministic and Randomized Multiparty Communication
LOCATION:32-G575
DESCRIPTION:Series: None\nSpeaker: Toni Pitassi\, U. Toronto\nHost: Madhu Sudan\, MIT CSAIL\nContact: Joanne Hanley\, 617 253 6054\, joanne@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nNumber-on-forehead communication protocols are a fascinating model of computation where k collaborating players are trying to evaluate a function\, f. The players are all-powerful\, but the input to the function is partitioned into k pieces\, of n bits each where the $i^{th}$ piece\, $x_i$\, is placed\, metaphorically\, on the forehead of player $i$. Thus each player sees all but the input that is on his/her forehead. In order to compute $f$\, the players communicate by writing bits on a shared blackboard and the complexity of the protocol is the number of bits that are communicated. This model is extremely important and has found applications in a surprising variety of areas\, including circuit complexity\, pseudorandomness\, and proof complexity.\n\nIn this model\, a protocols is said to be efficient if it has complexity polylogn. Correspondingly\, $P^{cc}_k$\, $BPP^{cc}_k$\, and $NP^{cc}_k$ are the NOF communication complexity analogs of the standard complexity classes. For example\, $RP^{cc}_k$ is the class of functions having efficient one-sided-error communication complexity\, and one of the most fundamental open problems is to separate these classes.\n\nIn this work\, we exhibit an explicit function that can be computed by a nondeterministic NOF protocol communicating $O(logn)$ bits but requiring nearly linear number of bits of communication for randomized NOF protocols with $k=\delta \log n$ players for any fixed $\delta <1$. Thus\, we separate $NP^{cc}_k$ from $RP^{cc}_k$.\n\nThis is joint work with Matei David and Emanuele Viola.\n
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080709T150000
DTEND;TZID=US/Eastern:20080709T153000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1911
SUMMARY:AnalogySpace: Reducing the Dimensionality of Common Sense Knowledge
LOCATION:MIT Media Lab\, 483A (Reef Room)
DESCRIPTION:Series: None\nSpeaker: Rob Speer\, Catherine Havasi\, and Henry Lieberman\, CSAIL\, Brandeis University and the Media Lab\nHost: \, \nContact: Rob Speer\, 617-253-0315\, rspeer@mit.edu\nRefreshment Time: \nRelevant URL: http://analogyspace.media.mit.edu/\nThis is a practice talk for AAAI.\n\nAbstract: We are interested in the problem of reasoning over very large common sense knowledge bases. When such a knowledge base contains noisy and subjective data\, it is important to have a method for making rough conclusions based on similarities and tendencies\, rather than absolute truth. We present AnalogySpace\, which accomplishes this by forming the\nanalogical closure of a semantic network through dimensionality reduction. It self-organizes concepts around dimensions that can be seen as making distinctions such as ``good vs. bad'' or ``easy vs. hard''\, and generalizes its knowledge by judging where concepts lie along these dimensions. An evaluation demonstrates that users often agree with the predicted knowledge\, and that its accuracy is an improvement over previous techniques.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080709T130000
DTEND;TZID=US/Eastern:20080709T140000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1917
SUMMARY:microRNA networks regulated during neural and embryonic stem cell differentiation
LOCATION:32G-575\, MIT Stata Center\, 5th floor conference ro
DESCRIPTION:Series: None\nSpeaker: Loyal Goff\, Rutgers Stem Cell Research Center\nHost: Manolis Kellis\, CSAIL-MIT\nContact: Sally Lee\, 3-6837\, sally@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nMicroRNAs represent a group of functional non-coding RNAs (ncRNA) with\na role in both translational repression and/or RNAi-mediated\ndegradation of specific target mRNAs. Reports indicate that microRNA\nregulation plays an important role in numerous cellular processes. We\nhave identified microRNAs regulated during induction of neurogenesis\nin an in vitro model\, and have correlated the expression of these\nspecific microRNAs to regulated mRNAs with the goal of describing\npotential functional interactions between the two sets of molecules.\nWe demonstrate the active role these regulated microRNAs play in\ninducing a neuronal phenotype in uncommitted neural precursor cells.\nAs a result of this analysis\, as well as additional studies of\nregulated mRNA transcripts during neurogenesis\, we have identified a\nregulated transcription factor\, Mef2C\, which is chromosomally adjacent\nto one of the regulated microRNAs. The unexpected appearance of this\ntranscription factor and its relatively unexplored role in neural\ndifferentiation suggest that Mef2C may play a role in this cellular\nprocess. This single genomic locus containing an induced microRNA\ninvolved in neuronal fate determination\, as well as Mef2C\, is\ntranscriptionally regulated during neuronal cell specification. In\naddition\, HDAC4\, a known repressor of Mef2C activity is validated as a\ntarget for miR-9 mediated post-transcriptional repression. We outline\nhere a putative network of interactions that arise from\ntranscriptional activation of this locus\, with the end result\ncontributing to the induction of a neuronal phenotype. In addition\, we\nwill describe initial results from our deep sequencing analyses of\nmicroRNA expression in differentiating human embyonic stem cell lines\nand the identification of novel microRNA sequences.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080710T110000
DTEND;TZID=US/Eastern:20080710T120000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1919
SUMMARY:Incremental Smoothing and Mapping
LOCATION:32-397
DESCRIPTION:Series: None\nSpeaker: Michael Kaess\, Georgia Tech\nHost: \, \nContact: Nicholas Roy\, x3-2517\, nickroy@mit.edu\nRefreshment Time: \nRelevant URL: \nAbstract:\nIn this talk I begin with results from my earlier work on 2D and 3D laser-based robot mapping as well as robot localization. From there\, the need for real-time performance leads to my dissertation work called incremental smoothing and mapping (iSAM). iSAM provides an efficient and exact solution to the full simultaneous localization and mapping (SLAM) problem based on an incremental matrix factorization and periodic variable reordering. iSAM also supports data association by providing efficient access to the estimation uncertainties. I present results from iSAM applied in various contexts including visual SLAM\, for which tracked features are provided by my real-time visual odometry implementation.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080716T140000
DTEND;TZID=US/Eastern:20080716T150000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1921
SUMMARY:Unique Games on Expanding Constraint Graphs are Easy
LOCATION:32-G575
DESCRIPTION:Series: None\nSpeaker: Alexandra Kolla\, UC Berkeley / Microsoft Research\nHost: Nick Harvey\, MIT CSAIL / Microsoft Research\nContact: Nick Harvey\, \, nickh@mit.edu\nRefreshment Time: \nRelevant URL: \nWe present an efficient algorithm to find a good solution to the Unique Games problem when the constraint graph is an expander. We introduce a new analysis of the standard SDP in this case that involves correlations among distant vertices. It also leads to a parallel repetition theorem for unique games when the graph is an expander.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080717T110000
DTEND;TZID=US/Eastern:20080717T120000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1918
SUMMARY:Feature Specific Imaging
LOCATION:32-D507
DESCRIPTION:Series: None\nSpeaker: Professor Mark A. Neifeld\, University of Arizona-ECE Department/Optical Sciences\nHost: Fredo Durand\, MIT - CSAIL\nContact: Bryt Bradley\, 617-253-6583\, bryt@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nCoffee will be served at 10:45am in 32-D507\, just preceding the talk.\n\n - - - - -\n\nFeature-specific imaging (FSI) is a technique by which optical \nmeasurements employing a non-traditional basis may be used to efficiently \nextract spatial\, temporal\, and/or spectral object information. \nBecause the measurement dimensionality of a FSI system is often much \nlower than the native dimensionality of the object space\, FSI is \nsometimes called compressive imaging. \n\nThis presentation will discuss several candidate optical systems \nfor FSI. The performance of FSI will be analyzed for both general \npurpose imaging (i.e.\, image reconstruction) and task-\nspecific imaging (e.g.\, target detection and/or tracking). FSI will \nalso be discussed as a convenient framework within which the \njoint optimization of optical and post-processing resources may \nbe undertaken.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080718T130000
DTEND;TZID=US/Eastern:20080718T140000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1922
SUMMARY:Perceptually-based scalable graphics
LOCATION:32-D507
DESCRIPTION:Series: None\nSpeaker: Kavita Bala\, Cornell University - Computer Science Department and Program of Computer Graphics\nHost: Fredo Durand\, MIT - CSAIL\nContact: Bryt Bradley\, 617-253-6583\, bryt@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nCoffee will be served at 12:45pm in 32-D507\, just preceding the talk.\n\n - - - - -\n\nSimulating the appearance of complex scenes faithfully and\nefficiently is a fundamental challenge in graphics. My\nresearch develops algorithms that scale to complex\nillumination and scenes\, by exploiting limitations in\nhuman perception. In this talk I will describe two\ncomplementary research goals. First\, we need to\nunderstand image fidelity: when is a rendered image good\nenough? We introduce a new appearance-based measure of\nimage fidelity called visual equivalence[SIG07\,SIG08] that\ngoes beyond pixel accuracy to better capture what graphics\npractitioners care about: preserving appearance in complex\nscenes. \n\nSecond\, we design perceptually-based algorithms that scale\nto complex scenes and complex illumination effects\, like\nmotion blur\, depth-of-field\, participating media\, and\nsubsurface scattering. I will describe our work on\nlightcuts for scalable final rendering [SIG05\,SIG06\,EG08]\nand matrix row-column sampling for lighting preview\n[SIG07\,EGSR08]. These approaches are complementary in\nperformance and can be applied across a range of\napplications such as cinematic relighting\, production\nrendering\, games\, cultural heritage\, and ecommerce\, among\nothers.\n\nJoint work with Adam Arbree\, James Ferwerda\, Milos Hasan\,\nFabio Pellacini\, Ganesh Ramanarayanan\, Edgar\nVelazquez-Armendariz\, and Bruce Walter. \n\n------------------------------------------------------------\nBio: Kavita Bala is an Assistant Professor in the Computer\nScience Department and Program of Computer Graphics at\nCornell University. She received her S.M. and Ph.D. from\nthe Massachusetts Institute of Technology (MIT)\, and her\nB.Tech. from the Indian Institute of Technology (IIT\,\nBombay). Bala specializes in computer graphics\, leading\nresearch projects in scalable rendering\,\nperceptually-based rendering\, interactive global\nillumination\, and image-based modeling and texturing. She\nhas co-authored the graduate-level textbook "Advanced\nGlobal Illumination" (A K Peters publisher\, second\nedition). In 2005 she co-chaired the Eurographics\nSymposium on Rendering (EGSR). Bala has received the NSF\nCAREER award\, Cornell's College of Engineering James and\nMary Tien Excellence in Teaching Award\, and Cornell's\nAffinito-Stewart award.\n------------------------------------------------------------
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080718T130000
DTEND;TZID=US/Eastern:20080718T140000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1927
SUMMARY:Learning to modularize
LOCATION:32G-575
DESCRIPTION:Series: None\nSpeaker: Dr. Sameet Sreenivasan\, Center for Complex Networks\, U. of Notre Dame\nHost: Manolis Kellis\, CSAIL-MIT\nContact: Sally Lee\, 3-6837\, sally@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nA large number of biological networks have been shown to possess a\nfunctionally modular architecture comprising of units or pathways that\nfunction independently of each other. The factors for the emergence of\nsuch modules is not conclusively known. While proposals have been\nmade for the evolutionary emergence of modularity\, it is far from\nclear that evolutionary adaptation is the sole mechanism that can lead\nto functional specialization. In this talk\, I will focus on the\nquestion as to whether learning algorithms can generate modularity in\na network. I will demonstrate through computational examples that\nlearning in multitasking networks can give rise to functionally\nmodular architectures. In this proof-of-principle study\, I will use a\nfeedforward neural network architecture as a prototypical example of a\nbiological network\, and show some conditions under which learning to\nperform multiple tasks can give rise to functional modules. In\naddition\, the work I present will also demonstrate that searching for\nstructural modules may not necessarily enable us to identify the\nfuncional modules in networks.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080723T150000
DTEND;TZID=US/Eastern:20080723T160000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1925
SUMMARY:Data-driven grasping and manipulation
LOCATION:32-G449 Patil/Kiva
DESCRIPTION:Series: None\nSpeaker: Nancy S. Pollard\, Computer Science Dept\, CMU\nHost: Tomas Lozano-Perez\, CSAIL\nContact: Tomas Lozano-Perez\, 617-253-7889\, tlp@csail.mit.edu\nRefreshment Time: 2:45PM\nRelevant URL: \nAbstract: Data captured from human performances of activities ranging from the everyday through the extraordinary has become widely accessible over the past 10 years. The ability to download or capture human motion\, analyze it in great detail\, and process it for use in real-time algorithms has led to new ways of thinking about character animation and robot control. However\, we do not yet know how to make the most effective use of this data. What is important about a given performance? How can it be modified to create realistic new scenarios? And what are the limits of this approach.can we ever create behavior that could be called dexterous from a collection of observed performances?\n\nIn this talk\, I will focus on the problem of creating dexterous grasping and manipulation behaviors from observed performances. I will discuss how my ideas have changed over the past decade\, as we have gone from the idea that the important aspect of a grasp is how it allows us to apply forces to an object through consideration of hand shape\, anatomical constraints\, and dynamic properties to the observation that grasps often involve preparatory sensing and manipulation actions which we have shown can reduce the effort needed to acquire an object. Results in computer animation and robot control\, as well as results from controlled human subjects experiments will be presented.\n\nShort bio: Nancy Pollard is an Associate Professor in the Robotics Institute and Computer Science Department at Carnegie Mellon University. She received her PhD in Electrical Engineering and Computer Science from the MIT Artificial Intelligence Laboratory in 1994\, where she performed research on grasp planning for articulated robot hands. Before joining CMU\, Nancy was an Assistant Professor and part of the Computer Graphics Group at Brown University. She received the NSF CAREER award in 2001 for research on 'Quantifying Humanlike Enveloping Grasps' and the Okawa Research Grant in 2006 for "Studies of Dexterity for Computer Graphics and Robotics."
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080725T123000
DTEND;TZID=US/Eastern:20080725T133000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1923
SUMMARY:Massively parallel systems and global optimisation
LOCATION:4-237
DESCRIPTION:Series: None\nSpeaker: DR. NARENDRA KARMARKAR\, Laboratory for Computational Mathematics\nHost: \, \nContact: Shirley A. Entzminger\, 617-253-4357\, daisymae@math.mit.edu\nRefreshment Time: \nRelevant URL: \n*******************************************************************************\nCOMPUTATIONAL RESEARCH in BOSTON SEMINAR\n\n\n[NOTE: Different location.]\n\nDATE: Friday\, JULY 25\, 2008\nTIME: 12:30 PM\nLOCATION: Building 4\, Room 237\n\n(Pizza and beverages will be provided.)\n\n\nTitle:MASSIVELY PARALLEL SYSTEMS AND GLOBAL OPTIMISATION\n\nSpeaker:DR. NARENDRA KARMARKAR\nLaboratory for Computational Mathematics\n\n\nABSTRACT:\n\nWe will briefly describe recent breakthrough in design of massively parallel systems based on insights derived from global optimization problems having multiple global optima. These designs include:\n\n- Physical design of the projective geometry machine using massively\n parallel quantum tunneling\, which can totally overcome obstacles of\n latency and bandwidth faced by contemporary designs. The new design can\n broaden applicability of massive multi-threading to large and very\n general classes of computational problems\, and can be implemented using\n already known fabrication techniques.\n\n- Design of multi-ported\, low latency\, secondary storage based on\n magneto-optics\, implementing shared memory directly at physical level\,\n providing a highly valuable feature for data bases and transactional\n memory.\n\n- Design of new high bandwidth switches required for next generation\n internet infrastructure.\n\n- Design of novel robots with large number of "electro-magnetic\n fingers" for placing atoms based on complex and sparse patterns of\n multiple global minima that are more general than regular periodic\n patterns achieved before using interference lithography.\n\n- Design of control systems whose stability analysis requires\n liapunov-like functions with multiple basins of attraction.\n\n- Design of phased-array radars in tera-hertz range.\n\n- Computational calibration of parameters occurring in empirical\n force fields\, whose values may be difficult to measure experimentally\n but can be reverse engineered from known structure of folded proteins.\n\nThis work involves integration of ideas\, concepts and processes from many fields:\n\n- Math (Optimization theory\, Discrete subgroups of lie groups)\n- CS (Parallel Architectures)\n- Physics (Path integrals\, Quantum tunneling\, Optics\, Electron Optics)\n- EE (CMOS & MEMS processes\, Field emission devices\, Control Theory)\n- Material Science (Generalized interference lithography)\n\n\n[NOTE: "Extended Abstract" is attached to e-mail and can be found on the CRiB websit at the end of end of Dr. KARMARKAR's abstract.]\n\n*******************************************************************************\n\nMassachusetts Institute of Technology\nCambridge\, MA 02139\n\n\nhttp://www-math. mit.edu/crib\nFor information on CRiB\, contact:\nAlan Edelman: edelman@math.mit.edu\nSteven G. Johnson: stevenj@math.mit.edu\nJeremy Kepner: kepner@ll.mit.edu\nPatrick Dreher: dreher@mit.edu
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080725T100000
DTEND;TZID=US/Eastern:20080725T110000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1929
SUMMARY:Increasing the Robustness of Networked Systems
LOCATION:32-G449 Patil/Kiva Conference Room
DESCRIPTION:Series: None\nSpeaker: Srikanth Kandula\, MIT-CSAIL\nHost: Dina Katabi\, MIT-CSAIL\nContact: Mary McDavitt\, 617-253-9620\, mmcdavit@csail.mit.edu\nRefreshment Time: 9:45AM\nRelevant URL: http://nms.lcs.mit.edu/~kandula\nWhat popular news do you recall about networked systems? You've\nprobably heard about the several hour failure at Amazon's computing\nutility that knocked down many startups for several hours\, or the\nattacks that forced the Estonian government web-sites to be\ninaccessible for several days\, or you may have observed inexplicably\nslow responses or errors from your favorite web site. Needless\nto say\, keeping networked systems robust to attacks and failures is an\nincreasingly significant problem.\n\nWhy is it hard to keep networked systems robust? We believe that\nuncontrollable inputs and complex dependencies are the two main\nreasons. The owner of a web-site has little control on when users\narrive; the operator of an ISP has little say in when a fiber gets\ncut; and the administrator of a campus network is unlikely to know\nexactly which switches or file-servers may be causing a user's\nsluggish performance. Despite unpredictable or malicious inputs and\ncomplex dependencies we would like a network to self-manage itself\,\ni.e.\, diagnose its own faults and continue to maintain good\nperformance.\n\nIn this talk\, I will present a generic approach to harden networked\nsystems. For well understood systems that need to respond rapidly to\nunpredictable inputs\, we design online solutions that re-optimize\nresource allocation as inputs change. For systems with complex\ndependencies\, we devise techniques to infer these dependencies from\npacket traces and build functional representations that facilitate\nreasoning about most likely causes for faults. We present a few\nsolutions\, as examples of this approach\, that tackle an important\nclass of network failures. Specifically\, we address (1) re-routing\ntraffic around congestion when traffic spikes or links fail in\ninternet service provider networks\, (2) protecting web-sites from\ndenial of service attacks that mimic legitimate users and (3)\ndiagnosing causes for performance problems in enterprises and\ncampus-wide networks. Through implementations\, simulations and\ndeployments\, we show that our solutions significantly advance the\nstate-of-the-art.\n\nThesis Committee:\nProf. Dina Katabi (Advisor)\nDr. Paramvir Bahl\, Prof. Hari Balakrishnan and Prof. John Guttag
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080728T143000
DTEND;TZID=US/Eastern:20080728T153000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1926
SUMMARY:Thesis Defense: Autonomous Problem Solving by Inductive and Deductive Program Synthesis
LOCATION:36-428
DESCRIPTION:Series: None\nSpeaker: Harold Fox\, CSAIL\nHost: Harold Fox\, CSAIL\nContact: Harold Fox\, 617-953-6189\, hfox@csail.mit.edu\nRefreshment Time: 2:15PM\nRelevant URL: \nAbstract:\nHow do people learn abstract concepts unsupervised?\nPsychologists broadly recognize two types of concepts\, declarative\nknowledge and procedural knowledge: know-what and know-how. While much\nwork has focused on unsupervised learning of declarative concepts\nas clusters of features\, there is much less clarity on the\nrepresentation for procedural concepts and the methods for learning\nthem. In this thesis\, I claim that programs are a good representation\nfor procedural knowledge\, and that program synthesis is a promising\nmechanism for procedural learning. Prior attempts at AI program synthesis have taken a\npurely deductive approach to building provably correct programs.\nThis approach requires many axioms and non-trivial\ninteraction with a human programmer. In contrast\, this thesis\nintroduces a new approach called SSGP (Sample Solve Generalize Prove)\, which combines\ninductive and deductive synthesis to autonomously synthesize programs\n with no extra knowledge outside of the program specification. The approach is to\ngenerate examples\, solve the examples\,\ngeneralize from the solutions\, and then prove the generalization\ncorrect.\n\nThis thesis presents two systems\, Spec2Action and\nHELPS. Given a logical specification\, Spec2Action determines the\nrelations to change to perform simple operations on data\nstructures. The main part of its task is to uncover the recursive structure of\nthe domain from the purely logical input spec.\nHELPS generates sequential programs with loops and branches using\nSTRIPS actions as the primitive\nstatements. It solves generalizations of classic AI tasks like BlocksWorld.\nThe two systems use SAT solving and other grounded reasoning\ntechniques to solve the examples and generalize the solutions. To\nprove the abstracted hypotheses\, the systems use a novel theorem prover for doing\nrecursive proofs without an explicit induction axiom.
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=US/Eastern:20080805T150000
DTEND;TZID=US/Eastern:20080805T160000
URL;VALUE=URI:http://www.csail.mit.edu/events/eventcalendar/calendar.php?show=event&id=1928
SUMMARY:Thesis Defense: Engineering RNA Logic with Synthetic Splicing Ribozymes
LOCATION:32-G449
DESCRIPTION:Series: None\nSpeaker: Austin Che\, CSAIL\nHost: Tom Knight\, CSAIL\nContact: Austin Che\, 617-253-5899\, austin@csail.mit.edu\nRefreshment Time: \nRelevant URL: \nReusable components\, such as logic gates and code libraries\, simplify\nthe design and implementation of electronic circuits and computer\nprograms. The engineering of biological systems would benefit also\nfrom reusable components. In this thesis\, I evaluate the utility of\nsplicing ribozymes for the biological engineer. Ribozymes allow the\nengineer to manipulate existing biological systems and to program\nsynthetic self-modifying RNA systems. In addition\, splicing ribozymes\nare easy to engineer\, malleable\, modular\, and scalable.\n\nI use the model ribozyme from Tetrahymena to explore the principles\nbehind engineering biological splicing systems in vivo. I show that\nthe core ribozyme is modular and functions properly in many different\ncontexts. Simple design rules based on base pairing and computational\nRNA folding can predict splicing efficiency in bacterial cells. To\ntest our understanding of ribozyme function\, I generated synthetic\nribozymes by manipulating the primary sequence while maintaining the\nsecondary structure with standard base pairing. Results indicate that\nour biochemical understanding of the ribozyme is accurate enough to\nsupport engineering.\n\nSplicing ribozymes can form core components in an all-RNA logic\nsystem. I developed functioning biological transzystors\, switches\nanalogous to electrical transistors\, that can be engineered to use any\ntrans-RNA as input and any RNA as output. Transzystors allow us to\nread RNA levels genetically\, thereby using any RNA as an input in an\nengineered system. In addition\, the ribozyme can write RNA using the\ntrans-splicing reaction. Trans-splicing provides an easy mechanism\nto hook into an existing biological system and patch its\noperation. Using trans-splicing ribozymes\, I developed simple\,\nmodular\, and scalable biological logic gates. The generality of these\nribozymes for a wide set of applications makes them promising tools\nfor synthetic biology.\n\nThesis committee:\nTom Knight (supervisor)\nDave Bartel\nGeorge Church\nDenny Freeman
END:VEVENT
END:VCALENDAR