Dr. A.D.J. van Dijk
Applied Bioinformatics group and the Mathematical and Statistical Methods group of Wageningen University & Research
J Schell seminar room
UGent-VIB Research Building FSVM
9052 Zwijnaarde (Gent)
The amount of genomic data is growing rapidly, but our understanding of how the encoded genes perform their biological role lags behind. To contribute to this understanding, our research aims to obtain predictive models to analyse the relation between genes and traits, with a central role for gene networks.
We developed a gene function prediction approach, which uses networks such as co-expression networks, allowing to mitigate the often-used hypothesis that sequence-similarity between genes in different species indicates involvement in similar biological processes. Applications of the method include prediction of yeast gene functions, analysis of human disease genes, and comparison between gene functions in different plants. Detailed analysis of the input network used, indicates that "pruning" of highly connected nodes in a network has a positive effect on prediction performance. This is related to the tendency of hubs to have lower similarity to their neighbours compared to less well connected nodes. Integration of Quantitative Trait Locus (QTL) and Genome Wide Association Study data with gene function predictions leads to improved prediction performance. Moreover, predicted gene functions can be used to prioritize candidate genes in QTL regions, i.e., to find the most likely causal candidate gene underlying the trait of interest.
In addition to predicting what function genes perform, networks also enable to analyse how genes perform their function. Towards that end, we have been working on combining static statistical network models with dynamic network models. Focusing on flowering time regulation, we developed a dynamic model for a network of eight transcription factors involved in this process. This small network model was connected to a much larger network model for the various upstream genes involved in receiving environmental and endogenous signals. Data available for these genes consists of large amounts of gene expression data, typically of a static nature. Our approach is to use for the upstream network a Bayesian Network model, which we connect with the dynamic model. We demonstrate the predictive power of this approach by comparison with mutant data. This shows that the intricate network regulating flowering time can be successfully modelled by our approach, helping to understand how various genes collaborate in order to perform their biological role.
Dr. A.D.J. (Aalt-Jan) van Dijk performed his PhD in computational structural biology at Utrecht University (2006). He then moved to plant bioinformatics and systems biology, joining the Applied Bioinformatics group and the Mathematical and Statistical Methods group of Wageningen University & Research. His research involves mathematical modelling and statistical analysis of gene networks, and bioinformatics analysis of protein and DNA sequences. This involves development of algorithms as well as application to various types of biological datasets. His current research focus is on gene function prediction.