Large-scale analysis of human gene expression variability associates highly variable drug targets with lower drug effectiveness and safety

Abstract

Motivation The effectiveness of drugs tends to vary between patients. One of the well-known reasons for this phenomenon is genetic polymorphisms in drug target genes among patients. Here, we propose that differences in expression levels of drug target genes across individuals can also contribute to this phenomenon. Results To explore this hypothesis, we analyzed the expression variability of protein-coding genes, and particularly drug target genes, across individuals. For this, we developed a novel variability measure, termed local coefficient of variation (LCV), which ranks the expression variability of each gene relative to genes with similar expression levels. Unlike commonly used methods, LCV neutralizes expression levels biases without imposing any distribution over the variation and is robust to data incompleteness. Application of LCV to RNA-sequencing profiles of 19 human tissues and to target genes of 1076 approved drugs revealed that drug target genes were significantly more variable than protein-coding genes. Analysis of 113 drugs with available effectiveness scores showed that drugs targeting highly variable genes tended to be less effective in the population. Furthermore, comparison of approved drugs to drugs that were withdrawn from the market showed that withdrawn drugs targeted significantly more variable genes than approved drugs. Last, upon analyzing gender differences we found that the variability of drug target genes was similar between men and women. Altogether, our results suggest that expression variability of drug target genes could contribute to the variable responsiveness and effectiveness of drugs, and is worth considering during drug treatment and development. Availability and implementation LCV is available as a python script in GitHub (https://github.com/eyalsim/LCV). Supplementary information Supplementary data are available at Bioinformatics online.

Variability in Gene Expression is Associated with Incomplete Penetrance in Inherited Eye Disorders

“…Two gene expression resources were used to determine variability in population mRNA levels for IED-implicated genes: the GTEx [9] and the Eye Genotype Expression (EyeGEx) [18] databases. The GTEx v8 dataset (accessed September 2019) consisting of 948 donors, and 17,382 samples from Variability in mRNA levels in the population was assessed for the entire GTEx v8 dataset using a method similar to that developed by Simonovsky et al and using parameters broadly in keeping with those used by the GTEx consortium [19]. First, raw gene read data were downloaded from the GTEx web server and normalized to obtain the same library size for every sample for each tissue using the trimmed mean of M-values (TMM) method implemented in edgeR [20].…”

“…First, raw gene read data were downloaded from the GTEx web server and normalized to obtain the same library size for every sample for each tissue using the trimmed mean of M-values (TMM) method implemented in edgeR [20]. Prior to normalization, genes with at most 10 reads in all samples were removed to exclude lowly expressed genes, as these were considered to represent noise [19]. A comprehensive list of protein-coding genes was obtained from Ensembl Biomart [21].…”

The international epidemiological transition and the education gender gap

“…These well-established gender differences in the effectiveness of vaccines against infectious diseases contrast with the evidence on drugs and other methods of disease control. In their analysis of the effectiveness of 113 drugs Simonovsky et al (2019) find no systematic differences between men and women. Only for drugs acting on the central nervous system, such as anti-psychotic drugs and antidepressants, as well as beta-blockers, reducing the heart rate and systolic blood pressure there is weak evidence for more beneficial effects in women compared to men (Franconi et al 2007).…”

Variability in gene expression is associated with incomplete penetrance in inherited eye disorders

“…Variability in mRNA levels in the population was assessed for the entire GTEx v8 dataset using a method similar to that developed by Simonovsky et al and using parameters broadly in keeping with those used by the GTEx Consortium [19]. First, raw gene read data were downloaded from the GTEx web server and normalized to obtain the same library size for every sample for each tissue using the trimmed mean of M-values (TMM) method implemented in edgeR [20].…”