Table 1. Summary of significantly interacting genes from the CGA-LMM model, simple linear regression, and the ConCensusGLM model (Johnson et al, 2019).
CGA-LMM | Linear Regression | ConCensus GLM | ||||
---|---|---|---|---|---|---|
Drug | Expected target | Number of significant genes (Zrobust < -3.5) | Rank of expected target (out of 152) | Number of genes with negative slope significantly different than 0 † (out of 152) | Number of significant genes** (out of 152) | Rank of expected target (out of 152) |
Trimethoprim | trpG * | 4 | #2 | 30 | 138 | #15 |
Methotrexate | trpG * | 4 | #1 | 15 | 145 | #3†† |
Rifampin | rpoB | 4 | #14 | 64 | 148 | #3††† |
BRD-4592 | trpA | 4 | (not in hypomorph library) | 37 | 152 | ------ |
*TMP and MTX bind to and inhibit DHFR. However, the Johnson et al (2019) study showed that trpG has a chemical-genetic interaction with antifolate drugs.
† test of t-statistic, adjusted P-value<0.05 using Benjamini-Hochberg procedure for multiple test correction
**significance defined as minimum p-value < 1e-10 over same concentrations as analyzed for CGA-LMM, and mean log-fold-change<0
†† tied with 6 genes with Pvalue = 0; trpG is ranked 3rd based on mean log-fold-change
††† tied with 49 genes with Pvalue = 0; rpoB is ranked 3rd based on mean log-fold-change