Fig. 3
From: ulfasQTL: an ultra-fast method of composite splicing QTL analysis
The diagram of ulfasQTL. The left panel is basic flowchart and the right panel shows an illustrative example of core steps. There are in seven steps in the flowchart and the 3 steps shown in purple are core steps of ulfasQTL. We have two matrices after data preprocessing. One for the genotype data (genotypes of each sample at loci a and b) and one for the isoform expression data (isoforms A1, A2, A3 of gene A and isoforms B1, B2 of gene B of each sample). Then isoform expression matrix is converted to splicing components matrix (ϕ A1, ϕ A2, ϕ B1 for each sample) by the spherical coordinate system. We then send genotype data and splicing components matrix into MatrixeQTL (ANOVA model) to get the F statistics (F A1,a , F A2,a , F A1,b , F A2,b , F B1,a , F B2,b ) for each component-variant pair. Finally, we apply statistical transformations and get corresponding Z statistics according to F statistics, and merge Z statistics into S statistics (S A,a , S A,b , S B,a , S B,b ) for each gene-variant pair