Improving Accuracy of Gene Prediction Programs of the G*

描述（由申请人提供） 将GeneMark家族的基因识别程序的基因预测推广到使用贝叶斯分割方法分析密切相关的基因组。 这项研究将主要在俄罗斯进行，作为NIH资助#5R01HG00783的扩展。 该项目的目的是建立两个进化上接近的基因组的全球基因组比对，并使用修改的分割算法来解析基因组序列的同线区域。为此，将构建基因组的延伸区域的链比对。比对序列对将被解析成具有不同序列变异统计的片段。将建立一些基因组比对的统计模型，并设计自动选择与特定区域比对相关的模型的系统。本项目的最终目标是开发一种新的基于相似性的基因发现算法，用于近缘物种DNA序列的成对比对。 所有软件程序都将通过万维网界面提供给用户。

DESCRIPTION (provided by applicant) Generalization of the GeneMark family of gene recognition programs of gene prediction to analysis of closely related genomes using Bayesian segmentation approach. This research will be done primarily in Russia as an extension of NIH grant #5R01HG00783. The aim of this project is to build a global genomie alignment of two evolutionarily close genomes and to use a modification of the segmentation algorithm to parse syntenic regions of genome sequences. To this end the chain alignment of the extended regions of genomes will be constructed. Pairs of aligned sequences will be parsed into segments with different sequence variation statistics. A number of statistical models of genomic alignments will be built and the system, which automatically chooses the model relevant for the alignment of the particular region, will be designed. The final objective of the project is to develop a new algorithms of similarity based gene finding in the pairwise alignments of DNA sequences of closely related species. All the software programs will be available for users via the WWW interface.