List of the primers used for the qPCR and qRT-PCR reactions. (XLSX 31 kb)

List of genes with significant differential expression between FCB and FCB pfmdr1_3â ˛KD in a stage-specific manner. Genes with significantly higher expression displayed absolute fold change values >1.5 and normalized fold changes values two standard deviations above the median normalized fold change value. Fold change for each gene at the designated stage is indicated. No gene appeared to be significantly differently expressed between the two strains during both ring and schizont stages. Genes from antigenic variant families have been removed. (XLSX 71 kb)

Gene expression correlation between P. falciparum genes and either pfcrt or pfmdr1. Pearson correlation coefficients were calculated between pfcrt/pfmdr1 expression values and those of other 2,600 P. falciparum genes across 110 pairwise comparisons. These correspond to bi-directional comparisons between 11 parasite transcriptome data sets, including FCB and 7G8 data sets. (A) List of genes whose expression is highly correlated (PCC >0.7) to that of pfcrt. (B) List of genes whose expression is highly anti-correlated (PCC 0.7) to that of pfmdr1. (D) List of genes whose expression is highly anti-correlated (PCC

Genes with significant differential expression between 7G8 pfcrt_CTL and 7G8 pfcrt_T76K in a stage-specific manner. Genes with significantly higher expression displayed absolute fold changes >1.5 and normalized fold changes values two standard deviations above the median normalized fold change value. Genes with significantly lower expression had absolute fold change values lower than 0.6 and a normalized fold change three standard deviations below the median value. Fold change for each gene at the designated stage is indicated. (XLSX 58 kb)

List of genes with significant differential expression between FCB and 106/1 in a stage-specific manner. Genes with significantly higher expression displayed absolute fold changes >1.5 and normalized fold changes values two standard deviations above the median normalized fold change value. Genes with significantly lower expression had absolute fold change values lower than 0.6 and a normalized fold change three standard deviations below the median value. Fold change for each gene at the designated stage is indicated. (XLSX 66 kb)

Identification of genes whose expression is strongly correlated to that of both pfcrt and pfmr1. PCC were estimated from 110 pairwise comparisons of 11 transcriptome data sets. (A) List of genes organized by functional categories whose expression is highly correlated (PCC >0.7) to that of both pfcrt and pfmdr1. *Indicates genes with confidence, as found when using PCC computed from 55 unidirectional pairwise comparisons of the 11 transcriptome data sets. (B) List of genes organized by functional categories whose expression is highly anti-correlated (PCC

Values for gene expression fold differences between strains in the 7G8 data set for genes on chromosome 2. The boxed area corresponds to the region on chromosome 2 containing genes with a significant fold change (values <0.6). (XLSX 13 kb)

List of genes within the pfmdr1-containing amplicon with associated fold change between FCB and both FCB pfmdr1_3â ˛KD and 106/1, as measured by microarrays. (XLSX 11 kb)

Hierarchical clustering, which was performed using Pearson correlation coefficients between ~2600 gene pairs across 55 pairwise comparisons of 11 parasite transcriptome data sets, including FCB and 7G8 data sets. Cluster 1 corresponding to gene pairs with high Pearson correlation (PCC >0.7), containing pfmdr1 and pfcrt as well as a transcription factor of the ApiAP2 family PF3D7_0516800. (XLSX 42 kb)

Validation of the data point imputation. Using the publically available transcriptome datasets for 3D7, Dd2 and HB3 from (Llinas et al. NAR 2007), we validated whether the density of the data in our time series (of eight points, at a 6-hour resolution) was sufficient to impute data to a 7.5Â min resolution. For each strain, we selected eight time points evenly spread (6-hrs apart) from which we generated an imputed curve of 384 points. Using Pearson correlation values for gene expression, we then compared each actual hourly time point to the imputed curve to identify the closest matching time point. The imputed percentage column shows where the match occurred in the time series. Actual time points from the original time courses match imputed time points that are fairly evenly spread along the imputation-derived curve. (XLSX 37 kb)