Predicting protein-protein interactions

We are developing computational tools to distinguish physical protein-protein interactions from interactions only representing functional relationships between genes. We have used those methods to predict physical protein associations in human and fission yeast.

Read more and access the data.

Transcription Factor Target Gene Identification

Predicting TF target genes from DNA binding data.

Whereas a lot of eff ort has been put into the measurement of TF binding events, prediction of TF targets from DNA binding data is still an unresolved problem. Yet, this last step is critical for the functional interpretation of the many TF ChIP-seq studies that are currently being conducted.

In the past we have developed methods for the improved prediction of TF target genes based on diverse experimental data.

Read more and download the TargetCaller R-package.

Random Forest QTL Mapping

Random Forest

Mapping quantitative traits using Random Forest. This method accounts for possible non-additive (epistatic) interactions between loci.