The Construction Of Improved Gene Annotation For Phytopathogens

Plant bacteria pathogens cause devastating losses in agriculture over the world. The progress in plant pathogen projects has enabled a better understanding of phytopathogens at the molecular level. Up to the end of 2007, sixteen strains of bacteria pathogenic for plants have been sequenced. However, due to the absence of abundant experimental evidence, many problems still existed in the annotations of the sequenced bacterial genomes in current public databases, especially in GC-rich genomes. Due to these problems, we construct a database of improved gene annotation in phytopathogens (DIGAP).Features of DIGAP are as follows. (i) The 'hypothetical genes' in each plant pathogens are re-identified. Some 'hypothetical genes' have been recognized as non-coding sequences by using Z curve method. Evidence of principle component analysis (PCA), average length distribution and COG functional category occupation show that the identified non-coding ORFs are very unlikely to encode proteins. (ii) The translation initiation sites (TISs) of all the protein-coding genes in these genomes have been refined based on the ProTISA database and an ab initio program, GS-finder. (iii) Potential functions of a large number of 'hypothetical genes' have been predicted by using sequence alignment tools. (iv) Two theoretical gene expression indices, CAI and E(g) values, are calculated to show the gene expression levels. (v) Orthologs of the antibacterial therapeutic drug targets in human are enumerated in the database, which will provide potential targets in these plant pathogens. All the refined information will provide more accurate annotation for the research of these phytopathogens and antibiotic discovery in plant protection.The construction of the database is based on LAMP platform. DIGAP can be freely available at http://sdbi.sdut.edu.cn/ DIGAP/.