Examining the annealing schedules for RNA design algorithm
| Autor: | Stas Kalashnikov, Sinem Sav, Herbert H. Tsang, Halid Emre Erhan |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2016 |
| Předmět: |
0301 basic medicine
Optimization Schedule Mathematical optimization Computer science Bioinformatics 0206 medical engineering Evolutionary algorithm Geometry 02 engineering and technology Evolutionary algorithms Secondary structures Evolutionary computation Simulated annealing Annealing 03 medical and health sciences Biological process Hamming distance Catalytic element Optimization techniques Nucleic acid structure Protein secondary structure Folded structures Problem solving biology Primary sequences Ribozyme RNA Adaptive cooling schedule 030104 developmental biology RNA Sequence biology.protein Heuristics algorithm Algorithm design Heuristics Cooling Algorithm 020602 bioinformatics |
| Zdroj: | IEEE Congress on Evolutionary Computation, CEC 2016 CEC |
| Popis: | Date of Conference: 24-29 July 2016 Conference name: IEEE Congress on Evolutionary Computation (CEC), 2016 RNA structures are important for many biological processes in the cell. One important function of RNA are as catalytic elements. Ribozymes are RNA sequences that fold to form active structures that catalyze important chemical reactions. The folded structure for these RNA are very important; only specific conformations maintain these active structures, so it is very important for RNA to fold in a specific way. The RNA design problem describes the prediction of an RNA sequence that will fold into a given RNA structure. Solving this problem allows researchers to design RNA; they can decide on what folded secondary structure is required to accomplish a task, and the algorithm will give them a primary sequence to assemble. However, there are far too many possible primary sequence combinations to test sequentially to see if they would fold into the structure. Therefore we must employ heuristics algorithms to attempt to solve this problem. This paper introduces SIMARD, an evolutionary algorithm that uses an optimization technique called simulated annealing to solve the RNA design problem. We analyzes three different cooling schedules for the annealing process: 1) An adaptive cooling schedule, 2) a geometric cooling schedule, and 3) a geometric cooling schedule with warm up. Our results show that an adaptive annealing schedule may not be more effective at minimizing the Hamming distance between the target structure and our folded sequence's structure when compared with geometric schedules. The results also show that warming up in a geometric cooling schedule may be useful for optimizing SIMARD. © 2016 IEEE. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|