Aarhus University Seal / Aarhus Universitets segl

Functional validation of candidate genes detected by genomic feature models

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaperJournal articleResearchpeer-review


  • 1659.full

    Final published version, 1.05 MB, PDF document


  • Palle Duun Rohde
  • ,
  • Solveig Østergaard
  • ,
  • Torsten Nygaard Kristensen, Section for Biology and Environmental Science, Department of Chemistry and Bioscience, Aalborg University
  • ,
  • Peter Sørensen
  • Volker Loeschcke
  • Trudy F C Mackay, Department of Biological Sciences and Program in Genetics, North Carolina State University, United States
  • Pernille Merete Sarup
Understanding the genetic underpinnings of complex traits requires knowledge of the genetic variants that contribute to phenotypic variability. Reliable statistical approaches are needed to obtain such knowledge. In genome-wide association studies, variants are tested for association with trait variability to pinpoint loci that contribute to the quantitative trait. Because stringent genome-wide significance thresholds are applied to control the false positive rate, many true causal variants can remain undetected. To ameliorate this problem, many alternative approaches have been developed, such as genomic feature models (GFM). The GFM approach tests for association of set of genomic markers, and predicts genomic values from genomic data utilizing prior biological knowledge. We investigated to what degree the findings from GFM have biological relevance. We used the Drosophila Genetic Reference Panel to investigate locomotor activity, and applied genomic feature prediction models to identify gene ontology (GO) cate- gories predictive of this phenotype. Next, we applied the covariance association test to partition the genomic variance of the predictive GO terms to the genes within these terms. We then functionally assessed whether the identified candidate genes affected locomotor activity by reducing gene expression using RNA interference. In five of the seven candidate genes tested, reduced gene expression altered the phenotype. The ranking of genes within the predictive GO term was highly correlated with the magnitude of the phenotypic consequence of gene knockdown. This study provides evidence for five new candidate genes for locomotor activity, and provides support for the reliability of the GFM approach
Original languageEnglish
JournalG3: Genes, Genomes, Genetics (Bethesda)
Pages (from-to)1659-1668
Number of pages10
Publication statusPublished - 5 May 2018

See relations at Aarhus University Citationformats

Download statistics

No data available

ID: 124215280