Fig. 3

Correlations between expression patterns of adjacent mRNA transcripts. a Schematic diagram of gene pairs in a “head-to-head” orientation (also known as divergent gene pairs). The black arrows represent the direction of transcription and the dark grey box between the genes represents the intervening genomic sequence that is between the longest detection version of both 5′ UTRs. b Correlation of gene expression (in TPM, using Pearson correlation) for 1119 pairs of head-to-head genes with annotated 5′ UTRs plotted against the distance of the intervening genomic sequence. The median intervening sequence length is 548 bp (without annotation of 5′ UTRs this distance was 1946 bp). The median correlation of expression was 0.49, with the distribution showing a positive skew. An individual region (blue) is shown in more detail in panel C. c Expression profiles through the 3D7 time course for a head-to-head gene pair (Pf3D7_1011900, heme oxygenase and Pf3D7_1012000, RING zinc finger protein, putative). The steady state levels of these two genes is tightly correlated, with an R value of 0.96 (measured by spearman correlation). d Schematic diagram of gene pairs in a “tail-to-tail” orientation (also known as convergent gene pairs). The black arrows represent the direction of transcription and the dark grey box between the genes represents the genomic sequence that is between the longest detection version of both 3′ UTRs. e Correlation of gene expression (in TPM, using Pearson) for 1059 pairs of tail-to-tail genes with annotated 5′ UTRs plotted against the distance of the intervening genomic sequence. The median intervening sequence length is − 124 bp, i.e. most 3′ UTR pairs overlap (without annotation of 3′ UTRs this distance is + 657 bp). The distribution of correlation values includes both strongly negative and strongly positive relationships. An individual region (blue) is shown in more detail in panel f. f Expression profiles through the 3D7 time course for a tail-to-tail gene pair (Pf3D7_1115900, DHHC9 and Pf3D7_1116000, RON4). Despite an overlap of 327 nt in the 3′ UTRs the steady state level of these genes is strongly correlated, with a spearman correlation of value 0.81. g Correlation of expression profiles (Spearman) of 1000 neighbouring gene pairs for head-to-head, tail-to-tail and randomly selected gene pairs. The 1000 random neighboring gene pairs were randomly sampled 1000 times from all annotated head-to-head, tail-to-tail and tandem gene pairs. Mean correlations were 0.35, 0.10, and 0.18 for head-to-head, tail-to-tail and random orientations, respectively. The Wilcoxon rank sum test was used to determine significance between groups. P-values of 2.2e-16 and 3.8e-3 when comparing head-to-head and tail-to-tail groups to random pairings, respectively. h Correlation of expression profiles (Spearman) neighbouring gene pairs for head-to-head gene pairs, binned by intervening genomic distance. i Correlation of expression profiles (Spearman) neighbouring gene pairs for tail-to-tail gene pairs, binned by intervening genomic distance