engleski

Semi-automatic genome scanning of bacteria to identify promising secondary metabolite biosynthesis clusters for agro-biological exploitation

Genome sequencing of bacteria is expanding rapidly with over 500 genome sequences publicly available. Current advances in sequencing technology will lead to an explosive increase in known genome sequences in the near future. Traditional screening approaches will miss many interesting compounds, because they are not produced under the conditions used. Additionally, purification of an unknown compound is costly and time-intensive. Secondary metabolite synthesis clusters can be recognised by bioinformatic methods and allow partial prediction of the nature of the compounds produced. We have developed a semi-automatic annotation program (ClustScan) that uses biochemical knowledge to make predictions of products. The program is particularly effective for modular polyketide synthesis (PKS) clusters, which account for many compounds of agricultural and medical interest. It predicts as far as possible the extension reactions in a cluster. ClustScan runs on a server with a Java client on the user's computer. It finds PKS genes and predicts the presence of domains using specially developed HMM-profiles. The AT-domains are classified into probable C2- and C3-incorporating domains using 30 diagnostic amino acid positions. The degree of reduction is predicted by searching for active reduction domains (KR, DH, ER). The chirality of the extender unit is predicted using diagnostic amino acid residues in KR. We shall demonstrate the use of the program to analyse actinomycete genomes and show the degree of prediction of chemical structures that is possible. The resulting structures can be screened in silico to identify those with promise of activity. The prediction can also be used to design optimal detection and purification strategies to isolate the compounds.

genomes; gene-clusters; annotation; domain prediction

nije evidentirano