The expression of causal genes and their paralogs is highly imbalanced in their disease tissues compared to unaffected tissues.

<p>(A) The ratio between the expression levels of causal genes and their paralogs across tissues. Each point represents a single pair; the ratio in the disease tissue appears in red, the median ratio in unaffected tissues appears in gray. Ratios in the disease tissue (DT) were significantly higher than in unaffected tissues (UAT) when all pairs were considered ('All pairs', Mann-Whitney test, p<10⁻¹⁵); when only causal genes with single paralogs were considered ('Single CGP', Mann-Whitney test, p = 5.8*10⁻³); and when causal genes with multiple paralogs were considered and ratio was computed against their combined expression ('Multiple CGP', Mann-Whitney test, p = 1.2*10⁻³). ** refers to p<10⁻³ and *** to p<10⁻⁵. (B)—(E) The ratios between the expression levels of causal genes and their paralogs shown separately for causal genes sharing the same disease tissue. Each point represents the ratio observed in the disease tissue (red) and in an unaffected tissue (gray). The panels show genes causal for diseases manifesting in the brain (B), heart (C), muscle (D), and skin (E). In all panels, the median ratio is highest for pairs in their respective disease tissues. Moreover, except for brain, the difference in the distribution of ratios between disease tissue and unaffected tissues was statistically significant (brain p = 0.29, heart p = 3*10⁻⁴, muscle p = 2.45*10⁻⁵, and skin p = 7.6*10⁻⁶, Kolmogorov-Smirnov test). Additional disease tissues appear in <a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1007327#pgen.1007327.s005" target="_blank">S2 Fig</a>. In contrast to causal genes and their paralogs, the general difference in expression patterns between paralogs is very small (<a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1007327#pgen.1007327.s008" target="_blank">S5 Fig</a>). (F) The median expression ratio of causal genes and their paralogs for different subsets of diseases. Each row in the balloon plot corresponds to the subset of diseases that manifest in the tissue designated on the left, and entries per row show the median expression ratio of the respective causal genes and their paralogs across the different tissues. The median expression ratios were normalized to the highest median expression ratio observed in that row (reflected by circle size and color, with maximum value of 1). For most disease subsets, the highest median expression ratios were obtained in the respective disease-manifesting tissue. For brevity, 26 tissues were shown; a view of all tissues appears in <a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1007327#pgen.1007327.s006" target="_blank">S3 Fig</a>.