MobilomeFINDER: web-based tools for in silico and experimental discovery of bacterial genomic islands

Abstract

MobilomeFINDER (http://mml.sjtu.edu.cn/MobilomeFINDER) is an interactive online tool that facilitates bacterial genomic island or 'mobile genome' (mobilome) discovery; it integrates the ArrayOme and tRNAcc software packages. ArrayOme utilizes a microarray-derived comparative genomic hybridization input data set to generate 'inferred contigs' produced by merging adjacent genes classified as 'present'. Collectively these 'fragments' represent a hypothetical 'microarray-visualized genome (MVG)'. ArrayOme permits recognition of discordances between physical genome and MVG sizes, thereby enabling identification of strains rich in microarray-elusive novel genes. Individual tRNAcc tools facilitate automated identification of genomic islands by comparative analysis of the contents and contexts of tRNA sites and other integration hotspots in closely related sequenced genomes. Accessory tools facilitate design of hotspot-flanking primers for in silico and/or wet-science-based interrogation of cognate loci in unsequenced strains and analysis of islands for features suggestive of foreign origins; island-specific and genome-contextual features are tabulated and represented in schematic and graphical forms. To date we have used MobilomeFINDER to analyse several Enterobacteriaceae, Pseudomonas aeruginosa and Streptococcus suis genomes. MobilomeFINDER enables high-throughput island identification and characterization through increased exploitation of emerging sequence data and PCR-based profiling of unsequenced test strains; subsequent targeted yeast recombination-based capture permits full-length sequencing and detailed functional studies of novel genomic islands.

Figure 1.

MobilomeFINDER island discovery strategy: (A) MAmP facilitates identification of strains rich in microarray-elusive novel DNA (6), (B) tRNAcc identifies putative GIs in complete and near-complete closely related genomes (3), (C) tRIP PCR permits high-throughput identification and characterization of new GIs in test strains which can then be systematically targeted for capture (15,16) and full-length sequencing. ‘Island probing’ refers to a technique of tagging GIs with a dual positive–negative selectable cassette to allow for subsequent marker rescue and/or GI deletion experiments to size and further characterize islands (15). The yeast recombinational method is based on an E. coli–yeast shuttle vector that is constructed to carry targeting sequence homologues to conserved sequences flanking the putative island (16, 28). See text for further details.

Figure 2.

The ArrayOme web-tool. (A) Test strain CGH data file (17); (B) microarray index file that maps array probes to template CDS; (C–D) genome sequence and the annotated CDS files for major and minor reference templates, respectively; (E) input screenshot; (F) virtual MVG map of test strain; (G) major reference template map showing inferred contigs and gaps. See text for details.

Figure 3.

Web-based tRNAcc analysis of two Klebsiella pneumoniae genomes. (A) IdentifyIsland inputs specifying the K. pneumoniae genome sequences and the coordinates of tRNA genes. (B) MGH 78578 chromosomal map with tRNA genes marked. (C) Kp342 contigs were compared with the MGH 78578 genome using PipMaker (31) and ordered by coordinates of best alignment to produce a ‘virtual’ Kp342 genome. (D) IdentifyIsland output listing boundary coordinates of putative islands. (E) DNAnalyser input interface. (F–H) DNAnalyser output data for the two K. pneumoniae genomes analysed. See text for details.