Spotted PCR-product array
We have produced DNA probes by PCR amplification for most of the open reading frames (ORF) predicted by the sequencing of Streptomyces coelicolor A3(2) at the Sanger Centre. Pseudogenes, tranposons and insertion elements have not been included on the array
A program was written that checked each candidate probe for potential cross-hybridisation using similarity searches against the rest of the genome, rejecting those that had significant Blast hits in favour of unique probes. Where closely related (paralogous) genes are present, and cross-hybridisation cannot be avoided, probes that will pick up multiple genes were allowed. In practice this enabled the program to select a single probe for close gene families, such as a group of cold shock proteins, and the duplicated genes of the terminal inverted repeats. We also took steps to avoid generating multiple PCR products exploiting the genome sequence.
- automatic primer selection utilised several standard programs
- PCR from M145 genomic DNA provided by the John Innes Centre
- probe length 150-500 bp (smaller for small ORFs)
- two rounds of PCR using respectively gene-specific and universal primers
- universal primers allow covalent attachment to slides and relatively low cost re-amplification
Schematic representation of PCR amplification. The gene-specific (dark blue) portions of the first round primers are tagged with universal sequences (light blue), which can then be used to amplify all probes in a further amplification step. A C6-NH2 linker present on the forward universal primer allows covalent attachment to aldehyde-coated glass slides.
- primer selection program - primer3
- adapted for high GC content
- cross-hybridisation predicted using Blast2.0.9 searches
- mispriming predicted using Fasta3.2 searches
- multigene probes allowed where a close paralogue exists; defined as > 80% identity across whole ORF
- scripted to select first best pair
Primer design schema. Potential probes and primer pairs were checked against sequence data to determine cross-hybridisation (6) and mispriming (7) respectively. Regions predicted to hybridise to another gene and primers that would give extra bands were excluded (9). Where closely related (paralogous) genes (3) are present, and cross-hybridisation cannot be avoided, probes that will pick up multiple genes were allowed. In practice this enabled the program to select a single probe for close gene families - such as a group of cold shock proteins or duplicated genes - while designing probes for unique regions of less similar genes.