针对利用经典的随机上下文无关文法(SCFG)等模型对RNA(R ibonucle ic ac id)二级结构进行预测时,存在计算复杂性问题,该文给出了RNA二级结构的“新二级结构单元标签”(N SSEL)表示,相应提出了一种新的RNA二级结构预测的神经网络方法。...针对利用经典的随机上下文无关文法(SCFG)等模型对RNA(R ibonucle ic ac id)二级结构进行预测时,存在计算复杂性问题,该文给出了RNA二级结构的“新二级结构单元标签”(N SSEL)表示,相应提出了一种新的RNA二级结构预测的神经网络方法。这种二级结构的N SSEL表示格式很容易转换成常用的CT格式。基于tRNA数据集的实验表明,在完全相同的训练与测试数据集下,该方法,较之性能最好的B JK与BK 2等SCFG模型,其预测精度与相关系数都有所提高,证明了所提方法的可行性与有效性。由于神经网络启发式方法不存在计算时间复杂性问题,因此可望将此法用于预测SCFG等算法难以处理的大于1 000个碱基的长RNA序列的折叠问题。展开更多
An increasing number of structural homology search tools, mostly based on profile stochastic context-free grammars (SCFGs) have been recently developed for the non-coding RNA gene identification. SCFGs can include sta...An increasing number of structural homology search tools, mostly based on profile stochastic context-free grammars (SCFGs) have been recently developed for the non-coding RNA gene identification. SCFGs can include statistical biases that often occur in RNA sequences, necessary to profile specific RNA structures for structural homology search. In this paper, a succinct stochastic grammar model is introduced for RNA that has competitive search effectiveness. More importantly, the profiling model can be easily extended to include pseudoknots, structures that are beyond the capability of profile SCFGs. In addition, the model allows heuristics to be exploited, resulting in a significant speed-up for the CYK algorithm-based search.展开更多
文摘针对利用经典的随机上下文无关文法(SCFG)等模型对RNA(R ibonucle ic ac id)二级结构进行预测时,存在计算复杂性问题,该文给出了RNA二级结构的“新二级结构单元标签”(N SSEL)表示,相应提出了一种新的RNA二级结构预测的神经网络方法。这种二级结构的N SSEL表示格式很容易转换成常用的CT格式。基于tRNA数据集的实验表明,在完全相同的训练与测试数据集下,该方法,较之性能最好的B JK与BK 2等SCFG模型,其预测精度与相关系数都有所提高,证明了所提方法的可行性与有效性。由于神经网络启发式方法不存在计算时间复杂性问题,因此可望将此法用于预测SCFG等算法难以处理的大于1 000个碱基的长RNA序列的折叠问题。
文摘An increasing number of structural homology search tools, mostly based on profile stochastic context-free grammars (SCFGs) have been recently developed for the non-coding RNA gene identification. SCFGs can include statistical biases that often occur in RNA sequences, necessary to profile specific RNA structures for structural homology search. In this paper, a succinct stochastic grammar model is introduced for RNA that has competitive search effectiveness. More importantly, the profiling model can be easily extended to include pseudoknots, structures that are beyond the capability of profile SCFGs. In addition, the model allows heuristics to be exploited, resulting in a significant speed-up for the CYK algorithm-based search.
