Skip to main content
Premium Trial:

Request an Annual Quote

Study Compares Abilities of Various Genetic Approaches to Identify Drug Targets

NEW YORK – Researchers compared many major gene prioritization approaches to uncover which can best identify approved drug targets.

In a study published in Cell Genomics on Thursday, researchers from Switzerland and The Netherlands compared how well different approaches such as genome-wide association studies (GWAS), quantitative trait loci (QTL) analyses, and exome sequencing, along with network diffusion analyses, could identify drug targets for 30 disease traits.

"The strength of our analysis lies in the side-by-side comparison of gene prioritization methods that individually have proven to be successful in identifying drug targets," corresponding author Zoltan Kutalik, a researcher at University of Lausanne, and colleagues wrote in their paper.

Although GWAS is considered the most useful method to find drug targets, according to the authors, other approaches, such as large-scale molecular quantitative trait loci (mQTL) datasets and whole-exome sequencing have also facilitated the discovery of disease mechanisms and the identification of potential new drug targets. While the GWAS and QTL-GWAS methods focus on common genetic variants, the exome method considers only rare variants with minor allele frequencies below 1 percent, the authors noted.

The researchers wanted to compare the each method in drug target prioritization, for which they used various publicly available datasets and calculated gene scores. Previously, genes prioritized by these approaches have been more likely to be the targets of US Food and Drug Administration-approved drugs.

Overall, they found drug targets to be enriched among the genes prioritized by all these approaches. Specifically, enrichment odds ratios for drug targets were 2.17, 2.04, 1.81, and 1.31 for the GWAS, eQTL-GWAS, exome, and pQTL-GWAS methods, respectively. After accounting for differences in sample size and the number of testable genes by each method, the authors concluded that GWAS outperformed eQTL-GWAS and pQTL-GWAS, but not the exome approach, in identifying drug targets.

]The researchers also noted that genes prioritized via exome analysis differed from those highlighted by GWAS or the QTL-GWAS approaches.

They additionally leveraged network connectivity to identify potential drug target genes in close proximity to disease genes with otherwise limited genetic support. The inclusion of network analysis further improved the various approaches' target prioritization.   

The researchers, however, noted limitations of their study. For instance, the STRING network used in their connectivity analysis has non-random missing of network edges and may have a biased network structure.

The authors also pointed out that they did not factor in the directionality of therapeutic and genetic effects or whether the drug is an agonist or antagonist. "Although found to be less performant than GWAS, QTL-GWAS methods have the advantage of specifying directionality, as opposed to gene scores from the GWAS approach, which ignores SNP effect directions," they added.

They also noted that their analysis compared methods using historical drug discovery data as the ground truth, even though these data are likely biased. G-protein-coupled receptors, for example, are targets of about a third of FDA-approved drugs, though other genes may also be effective targets — but they just have never been tested in clinical trials, they said.

"Thus, our results may not reflect how well the tested genetic approaches uncover true disease genes, but rather how well they identify targets that were historically prioritized in drug development processes," they cautioned.