Department of Genetics, Sofia University "Saint Kliment Ohridski", Sofia, Bulgaria.
Comput Biol Med. 2022 Jun;145:105469. doi: 10.1016/j.compbiomed.2022.105469. Epub 2022 Apr 6.
Designing oligonucleotide-sensing ribozymes using computational approaches is advantageous to in vitro selection methods for efficiency and accuracy. Allosteric ribozymes can be computationally designed for various applications in gene therapy, designer gene control systems, biosensors, and molecular computation. Here we present two programs, the allosteric Ribozyme Generator (RG) and the Inverse Folding Ribozyme Generator (IFRG), engineered to generate allosteric ribozymes with YES logic. The RG computes allosteric ribozyme sequences' secondary structure using the minimal sequence of the hammerhead ribozyme by inserting oligonucleotide binding site (OBS) elements in the second stem. The IFRG program uses inverse folding to generate allosteric ribozyme sequences with OBS bearing distinct sequences and similar folding. For the generation of the OBS sequences, random search algorithms are employed. Allosteric ribozyme sequences generated by the RG can be used as a matrix for the IFRG program. This approach applies RNA-folding algorithms based on applying thermodynamic parameters using the partition function of the RNAfold, and the RNAinverse source codes from the Vienna RNA folding package. The two algorithms apply dynamic programming and random search algorithms to generate in silico allosteric ribozymes with predefined properties within minutes using a personal computer with over 90% accuracy, without high computation power as experimentally validated and published by us previously.
使用计算方法设计寡核苷酸感应核酶在效率和准确性方面优于体外选择方法。变构核酶可以通过计算设计用于基因治疗、设计基因控制系统、生物传感器和分子计算等各种应用。在这里,我们介绍了两个程序,变构核酶生成器(RG)和反向折叠核酶生成器(IFRG),它们被设计用于生成具有 YES 逻辑的变构核酶。RG 通过在锤头核酶的最小序列中插入寡核苷酸结合位点(OBS)元件,使用二级结构计算变构核酶序列的二级结构。IFRG 程序使用反向折叠生成具有不同序列和相似折叠的 OBS 承载的变构核酶序列。对于 OBS 序列的生成,使用随机搜索算法。RG 生成的变构核酶序列可以用作 IFRG 程序的矩阵。该方法应用基于 RNA 折叠算法的热力学参数,使用 RNAfold 的分区函数和维也纳 RNA 折叠包的 RNAinverse 源代码。这两个算法应用动态规划和随机搜索算法,在个人计算机上在几分钟内生成具有预定义属性的虚拟变构核酶,准确率超过 90%,而无需像我们之前实验验证和发表的那样需要高计算能力。