van Batenburg F H, Gultyaev A P, Pleij C W
Institute for Theoretical Biology, Leiden, The Netherlands.
J Theor Biol. 1995 Jun 7;174(3):269-80. doi: 10.1006/jtbi.1995.0098.
The possibilities of using a genetic algorithm for the prediction of RNA secondary structure were investigated. The algorithm, using the procedure of stepwise selection of the most fit structures (similarly to natural evolution), allows different models of fitness or driving forces determining RNA structure to be easily introduced. This can be used for simulation of the RNA folding process and for the investigation of possible folding pathways. Such an algorithm needs several modifications before it can predict RNA secondary structures. After modification, a fair number of correct stems are predicted, even when using computationally quick, but very crude, fitness criteria such as stem length and stacking energy, including elements of tertiary structure (pseudoknots). The fact that genetic algorithm simulation includes both stem formations and stem disruption allows one to observe intermediate structures that may be used in combination with phylogenetic or experimental research.
研究了使用遗传算法预测RNA二级结构的可能性。该算法采用逐步选择最适合结构的过程（类似于自然进化），使得能够轻松引入不同的适合度模型或决定RNA结构的驱动力模型。这可用于模拟RNA折叠过程以及研究可能的折叠途径。在该算法能够预测RNA二级结构之前，需要进行一些修改。修改后，即使使用计算速度快但非常粗略的适合度标准，如茎长度和堆积能量，包括三级结构元素（假结），也能预测出相当数量的正确茎。遗传算法模拟既包括茎的形成也包括茎的破坏，这一事实使人们能够观察到可与系统发育或实验研究相结合使用的中间结构。
