Unremarkable Problem, Remarkable Technique Operations in a DNA Computer DNA : A Unique Data Structure ! Pros and Cons Steps of His Experiement Major Breakthrough : Adleman’s Experiment DNA vs Silicon Conclusion – What does the future hold ?
1994, Leonard M. Adleman solved: An unremarkable problem, A remarkable technique The Problem : Hamiltonian Path Problem The Significance: Use of DNA to solve computation problems Computation at molecular levels ! DNA as a data structure ! Massively Parallel Computation
DNA Structure Double-stranded molecule twisted into a helix Sugar Phosphate backbone Each strand connected to a complimentary strand Bonding between paired nucleotides : Adenine and Thymine, Cytosine and Guanine Data Storage Data encoded as 4 bases : A,T,C,G Data density of DNA One million Gbits/sq. inch ! Hard drive : 7 Gbits per square inch Double Stranded Nature of DNA Base pairs – A and T, C and G S is ATTACGTCG then S' is TAATGCAGC Leads to error correction !
DNA Non Von Neuman, stochastic machines ! Approach computation in a different way Performance of DNA computing Affected by memory and parallelism Read write rate of DNA – 1000 bits/sec Silicon Von Neumann Architecture Sequential : "fetch and execute cycle" “the inside of a computer is as dumb as hell, but it goes like mad!” Richard Feynman
DNA Operations Fundamental Model Of computation : Apply a sequence of operations to a set of strands in a test tube Extract, Length, Pour, Amplify, Cut, Join, Repair, and many others ! Many copies of the enzyme can work on many DNA molecules simultaneously ! Massive power of DNA computation : Parallel Computation CPU Operations Addition, Bit-Shifting, Logical Operators (AND, OR, NOT NOR)
Leonard Adleman of the University of Southern California shocked the science world in 1994 He solved a math problem using DNA – The Hamiltonian Path Problem – Published the paper “Molecular Computation of Solutions of Combinatorial Problems” in 1994 in Science The field combines computer science, chemistry, biology and other fields. Prompted an "explosion of work," David F. Voss, editor of Science magazine
Exhaustive Search Branch and Bound 100 MIPS computer : 2 years for 20 cities ! Feasible using DNA computation 10^15 is just a nanomole of material Operations can be done in parallel Example Problem
Generate all the possible paths and then select the correct paths : Advantage of DNA approach Select paths that contain each city only once Steps taken by Adleman Select paths with the correct number of cities Select paths that start with the proper city and end with the final city Generate all possible routes
Strategy Encode city names in short DNA sequences. Encode paths by connecting the city sequences for which edges exist. Process ( Ligation Reaction ) Encode the City Encode the Edges Generate above Strands by DNA synthesizer Mixed and Connected together by enzyme - ligase
Random routes generated by mixing city encoding with the route encoding. To ensure all routes, use excess of all encoding ( 10 13 strands of each type ) Numbers on our side (Microscopic size of DNA) After This Step We have all routes between various cities of various lengths
Process (Polymerase Chain Reaction) Allows copying of specific DNA Iterative process using enzyme Polymerase Working : Concept of Primers Use primers complimentary to LA and NY Strategy Copy and amplify routes starting with LA and ending with NY After this Step Have routes of various lengths of LA….NY
Process (Gel Electrophoresis) force DNA through a gel matrix by using an electric field gel matrix is made up of a polymer that forms a meshwork of linked strands Strategy Sort the DNA by length and select chains of 5 cities After This Step Series of DNA bands –> select DNA with 30 bases
Process (Affinity Purification) Attach the complement of a city to a magnetic bead Hybridizes with the required sequence Affinity purify five times (once for each city) Strategy Successively filter the DNA molecules by city, one city at a time End result Path which start in LA, visit each city once, and end in NY
Alternate Method : Graduated PCR Series of PCR amplifications done Primer corresponding to LA and one other city Measure length of sequence for each primer pair Deduce position of city in the path One Method Simply sequence the DNA strands
Speed 10 14 operations per second 100x faster than current supercomputers ! Energy Efficiency 2 x 10 19 operations per joule. Silicon computers use 10 9 times more energy ! Memory 1 bit per cubic nanometer 10 12 times more than a videotape !
Amount Scales Exponentially For a 200 city HP problem, amount of DNA required > Mass of earth ! Stochastically driven process -> high error rates Each step contains statistical errors Limits the number of iterations
Current Trends Richard Lipton, Georgia Tech Surface DNA Techniques – U of Wisconsin 2010 – The first DNA chip will be commercially available Huge advances in biotechnology DNA sequencing Faster analysis techniques : DNA chips DNA : Molecule of the century Might be used in the study of logic, encryption, genetic programming and algorithms, automata, language systems.
Molecular Computation of Solutions to Combinatorial Computing Problems Leonard M. Adleman, Department of Computer Science, University of Southern California, 1994 On the Computation Power of DNA Dan Boneh, Christoper Dunworth, Richard J. Liption Department of Computer Science,Princeton University 1996 DNA Computing : A Primer Will Ryu