1 / 12

Stochastic Local Search Algorithms for DNA word Design

Stochastic Local Search Algorithms for DNA word Design. Dan C Tulpan, Holger H Hoos, and Anne Condon Summerized by Ji-Eun Yun. Introduction(1). Present results on the performance of a stochastic local search algorithm for the design of DNA code . Motivate

Download Presentation

Stochastic Local Search Algorithms for DNA word Design

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Stochastic Local Search Algorithms for DNA word Design Dan C Tulpan, Holger H Hoos, and Anne Condon Summerized by Ji-Eun Yun

  2. Introduction(1) • Present results on the performance of a stochastic local search algorithm for the design of DNA code. • Motivate • The tasks of storing information in DNA strands (for computation, molecular bar-codes) • Good word design  minimize error  higher information density • SLS Algorithm use randomised decisions while searching for solution to a given problem.

  3. Introduction • Focus • Finding codes of size greater than the best previously known bounds. • Applied to the design of DNA word sets. • Goal • To understand what Algorithmic principles are most effective in the application of SLS methods to the design of DNA or RNA word sets.

  4. overview • Introduction • Problem description • Constraint : Hamming distance, GC content, Reverse complement HD • SLS Algorithm detail • Result

  5. Problem Description(1) • A target k • Word length n • Find a set of k DNA words, each of length n, satisfying certain combinatorial constraint • Hamming Distance Constraint(HD) : • word w1, w2 • H(w1, w2) d : number of position i at which the ith letter in w1 differ from the ith letter in w2 • GC Content Constraint(GC) : a fixed percentage of the nucleotides within each word is either G or C(50%)

  6. Problem Description(2) • Reverse Complement Hamming Distance Constraint(RC) • w1 = w2 • H(w1,wcc(w2)) d • wcc(w2) : the watson-crick complement of DNA word w, obtained by reversing w and then by replacing each A in w by T(vice versa) and each C in w by G(vice versa) • Ex) ATGC  CGTA  GCAT • Use simple constraint • In order to understand whether our algorithmic techniques can find large sets of words.

  7. A Stochastic Local Search Algorithm(1) • SLS alg performs a randomised iterative improvement search in the space of DNA word sets. • Have exactly the target number of words • Always have the prescribed GC contents. • HD and RC constraints

  8. A Stochastic Local Search Algorithm(2) • Initial set of word is determined by a simple randomised process. • Generates any DNA word of length n with equal probability. • If GC is used, we check each word and only accept words designs use a GC contents close to 50%. • The possibility to initialise the search with a given set of DNA word • A set has less than k words, it is expanded with randomly generated. (the initial word set may contain multiple copies of the same word)

  9. A Stochastic Local Search Algorithm(3) Procedure Stochastic Local Search for DNA Word Design input: Number of words(k), word length(n), set of constraints(C) output: Set S of m words that fully or partially satisfies C for i := 1 to maxTries do S := initial set of words S’ := S for j := 1 to maxSteps do if S satisfies all constraints then return S end if Randomly select words w1, w2  S that violate one of the constraints M1 := all words obtained from w1 by substituting one base M2 := all words obtained from w2 by substituting one base with probability do select word w' from M1M2 at random otherwise select word w' from M1M2 such that number of conflict violations is maximally decreased end with probability if w'  M1 then replace w1 by w' in S else replace w2 by w' in S if S has no more constraint violations than S’ then S’ := S; end if end for end for return end Stochastic Local Search for DNA Word Design

  10. Result(1) • Run-time distributions • Run time Measured in terms of search step • CPU time per search step measured to obtain a cost model of these seach steps • Optimised parameter settings • DNA word sets of maximal size for various word lengths and combinatorial constraints.

  11. Result(2) • RTDs • MaxTries = 1000 • Without using random restart • Quality of word Sets obtained by our algorithm • Word length: 4~ 12 • HD only - significant body :16 best construction • HD+RD – 41case • HD+GC –almost all case • HD+RD+GC – Frutos(108), random initial set(92/108),Frutos set(109/108)

  12. Conclusion • Find high-quality sets of DNA words • Future work – more complex SLS - more diverse neighbourhood - other constraint(thermodynamic model)

More Related