Fig 1.
ICA is used to decompose the gene expression matrix Y into an IC coefficient matrix A and a component matrix S. Associations between the genotypes and coefficients in matrix A are tested to label any candidate genetic effects to be removed from the correction. In the example above, the first IC, shown in red, is marked as a candidate genetic component and the corresponding columns of A and rows of S are removed. Using the lower rank A* and S*, expression values originating from non-genetic components are reconstructed in Y*. Finally, K is created by calculating the sample covariance matrix of Y*, and included as a random effect in the mixed model for eQTL analysis.
Table 1.
Sample size for each subset of MuTHER dataset analyzed.
Table 2.
Sample size for each blood dataset analyzed.
Table 3.
Sample size for each GTEx dataset analyzed.
Fig 2.
Comparison of method performance for simulated data in the presence of sparse confounding factors.
(A) The recovery rate of simulated cis- (left), trans- (middle) and broad impact eQTL (right) for a range of FDR significance thresholds for each method averaging over the 50 simulated datasets with sparse confounding factors. The theoretical maximum recovery (TMR) shows the recovery when no confounding factors are included. (B) The Area Under the Curve (AUC) for the receiver operator characteristic (ROC) curves. (C) Box-plots of genomic inflation factors calculated for each method across the 50 simulated datasets with sparse factors.
Fig 3.
Significant eQTL discovered in MuTHER datasets for varying FDR thresholds.
Plots showing the counts of cis- and trans-eQTL versus a range of FDR significance thresholds for each of the methods applied to every dataset.
Fig 4.
Replication ratio of cis- and trans-eQTL in MuTHER and GTEx dataset pairs by each method.
The replication ratio calculated separately for cis- and trans-eQTL. The number of replicating eQTL are divided by the union of identified unique eQTL for each method in analyzing the (A) MuTHER twin pairs and (B) GTEx Tissue pairs.
Fig 5.
Replicating broad impact eQTL identified in the MuTHER LCL dataset.
(A) Chromosomes are plotted in the outermost circles with replicating broad impact trans-eQTL as blue bands in the inner layer, where red lines connect each trans-eQTL to the associated gene. (B) The number of replicating broad impact eQTL found in the MuTHER LCL twin pair by each method.