Streamlined representational difference analysis for comprehensive studies of numerous genomes

Abstract

Representational Difference Analysis (RDA) is a powerful but also quite complex and difficult method to reveal the variable genes among highly similar genomes or transcriptomes. This report is on the substantial simplification of RDA for microdiversity studies on Bacillus isolates. Following improvements are demonstrated: the average fragment size has been pushed up (∼700 bp) by using nucleotide-biased restriction enzymes, in this case the “high-GC” enzyme HhaI for low GC Gram-positive Bacillus. A PCR-based production of the DRIVER has been applied to avoid subculturing of strains and repeated large-scale DNA extraction. Tedious estimation and adjustment of DNA concentrations for the subtractive hybridization step were by-passed by coupling DRIVER and TESTER production. The problem of linear amplification of TESTER/DRIVER hybrids has been eliminated by applying two different adapters, one for the TESTER but also another one for the DRIVER. Hence, only one RDA round is required to achieve clear bands without background. After doing this streamlined DRIVER and TESTER production, the RDA analysis may now be performed in one night and day. This simplification will allow to scale up the number of genomes under investigation.