牙龈卟啉单胞菌编码基因重注释研究

Re-annotation of Porphyromonas gingivalis coding-sequences

为了确保牙龈卟啉单胞菌生物大分子信息的准确性,对NCBI数据库中的3株牙龈卟啉单胞菌的注释信息进行研究。首先,准备好蛋白质编码与非编码序列正负样本,用基于Z曲线理论的Fisher判别法对正负样本集进行训练,确定一个判断ORF编码或非编码的阈值t0,由阈值作为判别条件来识别所有的ORFs,判断基因片段是否具有编码蛋白质的功能,由此阈值为判别标准排除掉3株牙龈卟啉单胞菌基因组中错误的基因注释信息。然后,用Prodigal基因预测软件对牙龈卟啉单胞菌进行基因预测,基因预测结果与原始功能已知基因进行比对,挑选出具有不同5’终端的ORFs,将这些具有不同5’终端的ORFs与功能已知的基因片段进行比对,找到重叠率小于20%的候选基因。最后,对这些候选基因用Blast进行序列比对找到满足条件的新基因,并为这些新基因添加功能注释信息。基于以上方法共排除了117个非编码的开放式阅读框,并找到了30个NCBI数据库中缺失的编码蛋白质的新基因。

To ensure accuracy of P.gingivalis biological macromolecules information,we investigated the annotations of the 3 P. gingivalis based on NCBI database. Firstly,we prepared protein-coding and non-coding sequences as positive and negative samples,respectively,and used Fisher Discriminant which was designed based on Z curve theory to determine the threshold t0,which was used as the criterion to determine whether the gene encoding the protein or not. We firstly excluded the wrong annotation information from three stains of P. gingivalis based on the threshold. Secondly,the P. gingivalis were predicted with the prodigal gene prediction software. We used the predicted genes compared to the original known-function genes and selected the ORFs with different 5＇terminals,identified the candidate genes with overlapping rate of less than 20% from the ORFs with different 5 ＇terminals.Finally,we used the sequence alignment software Blast to find the candidate genes that meet the conditions. We excluded 117 non-coding open reading frames,and found 30 new protein-coding genes that were not annotated in the NCBI database.