Research

My main research interest is representation learning for biological data with an emphasis on single-cell (multi)-omics. Particularly, I am interested in how prior biological information can be encoded into such models and in identifying shared and complementary signals among different data modalities and exploiting them to build useful predicting models.

Some of my current projects are:

• Using artificial intelligence to combine multiple -omics measurements (mutations, RNA, methylation, metabolomics) from the blood of cancer patients to recommend a personalized treatment strategy. This involves developing novel algorithms to compare and prioritize the different measurements into a tool that will be tested on real clinical data. (with Amsterdam UMC)

• Probabilistic models to integrate single-cell RNA-seq data with whole genome sequencing of circulating DNA for minimally-invasive cancer detection. These DNA sequencing data contain (noisy) information about gene expression which can be used to detect tumor-derived fragments. The aim is to detect these fragments guided by cell-type-specific gene expression information.

• Representation learning of single-cell multi-omic data to uncover immune cell dynamics in response to vaccination. (with Sanquin, Dutch Blood Bank).

Internship projects for all topics above are available, for BSc and/or MSc students in bioinformatics, biomedical sciences, biological sciences, and computer science.