Stochastic versus stable transcriptional differences on Plasmodium falciparum DNA microarrays

The recent availability of the Plasmodium falciparum genome sequence has opened up convenient, large-scale analysis of transcriptional products in malaria. Protocols for cDNA labelling, cDNA hybridisation, and fluorescent signal detection developed for other organisms can be applied directly to malaria. However, P. falciparum offers unique challenges in data analysis due to stochastic variability in expression of some gene products, such as variable erythrocyte surface proteins. Careful comparison of global transcriptional patterns in two well-studied clones of P. falciparum (Dd2 and HB3) indicates that reliable, stable transcriptional alterations in malaria can be readily distinguished from stochastic processes. To do this, we utilised a complex experimental design which involves a combination of self-hybridisations and cross-hybridisations between two independently grown parasite populations for each clone being examined (for short, we call this a '2x2 CombiScan'). While even a simple 2x2 CombiScan required 12 microarray hybridisations, the effort generated output that was highly interpretable. Reliable RNA transcriptional differences between Dd2 and HB3 could be readily visualised using public algorithms for data normalisation and clustering.