The research line that I have developed over the last five years develops and applies computational algorithms to functional genomics datasets. The foundation for this work was laid during my PhD (2008, cum laude). While engaged in the first genome-wide association studies (GWAS), I concluded that, while knowledge on the individual genetic variants associated to disease provides insight into the determinants of disease, GWAS do not immediately provide insights into the mechanisms that drive disease.

My group works on multi-omics data generation and analysis, with a particular interest in the development of computational methods to identify the downstream molecular effects of these disease-associated genetic variants.

One strategy to do this is by performing expression quantitative trait locus (eQTL) mapping that permits the identification of the local (cis) and distal (trans) effects of genetic variants on gene expression levels. To identify these effects, I initiated a large blood eQTL meta-analysis consortium (eQTLGen) that revealed both cis- and trans-eQTL effects for many genetic risk factors (Westra et al,Nature Genetics, 2013, Zhernakova et al, Nature Genetics 2017). We have also systematically ascertained how genetic variation impacts methylation levels (Bonder et al, Nature Genetics2017), observing that genetic variation, gene expression and methylation levels are strongly correlated, and that methylation levels are indicative of transcription factor binding.

Another focus has been the development of novel methods to reuse publicly available data. I integrated gene expression data from 80,000 microarrays to accurately predict gene functions and gain better insight into somatic mutations in cancer (Fehrmann et al., Nature Genetics 2015). My group also developed DEPICT (Pers et al, Nature Communications 2015), which uses these predicted gene functions to better interpret GWAS findings. We recently re-did this analysis using 31,000 publicly available RNA-seq samples (Deelen et al,Nature Communications 2019) to make functional inferences about non-coding genes that are usually not captured well on microarrays. We also used this method to examine the clinical symptoms that mutations in these genes might cause and used these predicted clinical features as a new algorithm to increase the diagnostic yield of clinical exome-sequencing.

My group is currently concentrating on integrating large-scale multi-omics datasets by conducting large-scale trans-QTL meta-analyses in >30,000 samples (Vosa et al, BioRxiv 2018) in conjunction with single-cell RNA-seq data (Van der Wijst et al, Nature Genetics 2018) with the principal aims to conduct eQTL meta-analysis and to reconstruct personalized regulatory networks that can be used to better understand cancer-associated genetic variants. To do this optimally, we have initiated the single-cell eQTL consortium (Van der Wijst, ArXiv 2019, https://eqtlgen.org/single-cell.html) where over 20 international research groups work towards a large-scale federated cell-type specific eQTL analysis in >3,000 samples and reconstruction of cell-type specific gene regulatory networks.