A current paper in Plant Methods assessed the pros and cons of two RNA labeling methods for AGRONOMICS1 tiling arrays, concluding that random priming is more suitable for organelle transcriptome analysis as it can label non-polyadenylated transcripts effectively. They also generated new TAIR-10 based CDF files, which can be used to re-analyse existing AGRANOMICS1 CEL files. The new CDFs can be accessed here.
First of all, the authors gave an overview of the AGRONOMICS1 tiling array. It contains all the probes from the traditionally used Affymetrix ATH1 array, but has additional probes which mean the AGRONOMICS1 array yields expression data for over 7000 more genes, around a third of the genome. 90% of annotated genes on the TAIR9 database are on the array. Mitochondrial and chloroplast genomes are completely represented, and sRNA, tRNA and miRNA can also be detected. The AGRONOMICS1 array has probes that represent both strands of the entire Arabidopsis genome, allowing epigenetic profiling. The quality is comparable to that of the ATH1 array.
Müller et al. compared the GeneChip© IVT express kit, an oligo-dT based RNA labeling technique, with the GeneChip© whole transcript (WT) Sense Target Labeling Assay which uses random hexamers tagged with T7 promotor sequences. Both kits are from Affymetrix, Santa Carla, CA.
The team worked on two technical replicates of leaf RNA sample and three technical replicates of RNA samples from a flower for each method. The technical replicates from both methods were reproducible, and signal log ratios were not very different. Raw signal ratios were not comparable between the two methods. Expression signals for long transcripts and organellar transcripts were considerably higher using the random primer method.
The oligo-dT based method worked well on short and medium length, polyadenylated transcripts but was inefficient at labelling very long or non-polyadenylated transcripts. Random priming was a more versatile method than the oligo-dT method of RNA labeling for AGRONOMICS1 array chips, presenting a much more complete picture which included chloroplast and mitochondria gene expression – transcripts from organelles are not usually polyadenylated.
The methods agreed with each other on fold change of the bulk of the transcriptome and were both reproducible, so if you are used to oligo-dT labelling and are not working on organelles, there is not much advantage in switching.
Marlen Müller, Andrea Patrignani, Hubert Rehrauer, Wilhelm Gruissem and Lars Hennig (2012) Evaluation of alternative RNA labeling protocols for transcript profiling with Arabidopsis AGRONOMICS1 tiling arrays. Plant Methods 8:18 doi:10.1186/1746-4811-8-18