Fig. 4

Relative independence in the genetic control of cis-sQTLs and cis-eQTLs. a Distribution of Pearson’s correlation values relating PSI with gene expression. Case: PCC was calculated for PSI and expression of the same gene. Control: PCC was calculated for PSI and expression of a randomly selected gene. b Venn diagram showing the overlap between genes with cis-sGenes and those with cis-eGenes. From c to e, Manhattan plots show GWAS results for splicing level [51] and overall expression level of each gene (bottom). For boxplots, the lower and upper horizon lines are minimum and maximum values, respectively; the boxes span from the first to third quartiles; and the center line indicates the second quartile (median). p-values were calculated by two-sided Student’s t-test. c Manhattan plot showing a gene (GhVTE5/GH_A05G2930) detected to have a cis-sQTL without effect on expression level. Box plots display the association of sSNP haplotype (A05:35150218; AA, GG, and AG) with splicing level as indicated by PSI score [51] and with total mRNA level (bottom). d Manhattan plot showing a gene (GhSQN /GH_A02G1850) detected to have a cis-eQTL without effect on splicing. Box plots display the association of sSNP haplotype (A02:106438154; AA, GG, and AG) with splicing level as indicated by PSI score [51] and with total mRNA level (bottom). e Manhattan plot showing a gene (GhFBP/GH_A04G1526) detected to have both a cis-sQTL and cis-eQTL. Box plots display the association of sSNP haplotype (A04:85359924; AA, GG, and AG) with splicing level as indicated by PSI score [51] and with total mRNA level (bottom)