Laboratory of Translational Genetics

Our laboratory is a technology-driven translational research group. We aspire to tackle important questions in oncology by translating genome-scale data sets into clinically applicable knowledge. The objectives of our research are to leverage the genetic and epigenetic annotation of tumors to improve cancer prevention, diagnosis, prognosis and therapy. Investigations are based on the application of cutting-edge genome sequencing technologies (Ilumina NovaSeq, HiSeq4000, etc.) and advanced bioinformatics, and on the seamless integration of genomic data sets with clinical and fundamental biological information to generate novel insights and useful biomarkers for the field of oncology.

Our lab is located in the University Hospital Campus, Leuven, Belgium. We have longstanding ongoing collaborations with most oncologists at the University Hospital Leuven, positioning our lab at the center of translational cancer research in Leuven. We conduct translational research in numerous clinical trials run by the oncologists.


Our work is currently focused on two main research topics in the field of cancer genomics.

- Characterization how the tumor microenvironment (TME) determines tumor behavior and response to response to anti-angiogenic and checkpoint immunotherapies, which are both frequently used in the clinic (alone or in combination). This exciting line of research bears a lot of translational potential, as managing the levels of hypoxia in solid tumors by anti-angiogenic strategies represents a major yet incompletely understood challenge, while the field of immunotherapy is rapidly expanding - yet desperately in need of biomarkers predicting clinical response. We are using the newest single-cell profiling technologies to study these tumor micro-environments at the greatest detail (before, during treatment and at resistance).

The use of plasma cell-free DNA (cfDNA) as a diagnostic test to characterize and monitor tumors non-invasively. Specifically, we are optimizing methods to reliably detect somatic mutations, copy number changes, but also changes in DNA methylation and nucleosome positions in cfDNA from cancer patients.

Single-cell RNA sequencing of TME during cancer immunotherapy

Cancer consists of (epi)genetically and phenotypically different populations of single cells. This intra-tumor cellular heterogeneity underlies therapeutic resistance to cancer treatment. As such, improving our understanding of intra-tumor cellular heterogeneity, both in cancer cells and their surrounding stromal cells down to the single-cell resolution is essential to combat therapeutic resistance and to develop more effective cancer drugs. While cancer scientists have long sought to study single cells, the means and tools at their disposal have been limited. Recently, the field has witnessed the rapid emergence of single-cell sequencing technologies, with great capacity to simultaneously characterize the genome, transcriptome and epigenome of individual cells.

TME tumor cell

We recently published the very first ‘atlas’ of cell phenotypes found in lung tumors (Lambrechts et al., Nature Medicine 2018). By subjecting almost 100,000 single cells isolated from lung cancer patients without dissociation bias to single-cell RNA-sequencing (scRNA-seq), we were able to unbiasedly assess cancerous cells and non-cancerous cells in tumors such as blood vessels, immune cells and fibrous cells.



We discovered that the TME is more heterogenous than anticipated. We identified 52 stromal cell types versus the dozen cell types already known to be present. These cells have never before been characterized in their native environments. Because we compared with the matching non-malignant lung samples, we were moreover able to observe how each cell type is altered by the tumor. Our data indicated that the stromal cells targeted by anti-angiogenic and immunotherapies are distinct from those residing in non-malignant tissues. Highlighting such tumor-specific changes and vulnerabilities enables a more rational design of novel therapies targeting the TME.

Cancer cell clusters

In addition, an outstanding research question is to what extent this heterogeneity is similar between cancers affecting different organs. Therefore, we have profiled 233,591 single cells from patients with lung, colorectal, ovary and breast cancer and constructed a pan-cancer blueprint of stromal cell heterogeneity using different single-cell RNA and protein-based technologies. We identified 68 stromal cell populations, of which 46 are shared between cancer types and 22 are unique. We also characterised each population phenotypically by highlighting its marker genes, transcription factors, metabolic activities and tissue-specific expression differences. Resident cell types are characterised by substantial tissue specificity, while tumor-infiltrating cell types are largely shared across cancer types. By applying the blueprint to melanoma tumors treated with checkpoint immunotherapy and identifying a naïve CD4+ T-cell phenotype predictive of response to checkpoint immunotherapy, we illustrated how it can serve as a guide to interpret scRNA-seq data. In conclusion, by providing a comprehensive blueprint through an interactive web server, we have generated a first panoramic view on the shared complexity of stromal cells in different cancers. Our single-cell blueprint can be visualised, analysed and downloaded from an interactive web server ( (Qian et al., Accepted in Cell Research, a preprint present on bioRxiv 2020.04.01.019646).  

Cancer immunotherapy using immune checkpoint blockade has created a paradigm shift in the treatment of advanced-stage cancers. In terms of lives saved and person-years restored, these therapies promise to be more significant than any other form of cancer treatment. Leading healthcare professionals anticipate that the future of cancer therapy lies in designing novel treatments that boost the body's natural ability to fight cancer. However, one of the major limitations of immune checkpoint blockade is that it provides durable clinical responses only in a relatively small fraction of patients. Indeed, only 40% of patients with advanced metastatic melanoma respond to nivolumab (anti-PD1) and/or ipilimumab (anti-CTLA4), while response rates for other types of cancers are even lower. In addition, although long-lasting anti-tumor responses were observed, disease relapse due to resistance frequently occurs. This clearly illustrates how the success of immune checkpoint blockade varies among different cancer types, and how, even within the same cancer type, responses may hugely differ between patients. Unfortunately, it is still not possible to predict upfront which patient will respond to immune checkpoint blockade.

Nowadays, together with the UZ Leuven oncologists, we exploit the innovative single-cell technologies (scRNA-seq, CITE-seq, spatial single-cell profiling, single-cell DNA-seq, single-cell alternative splicing and single-cell CNV profiling) and apply multi-omics single-cell profiling in the context of clinical trials involving checkpoint immunotherapy. Specifically, tumor biopsies will be collected in patients receiving checkpoint inhibitors before and during treatment, as well as at disease progression. This will allow us to monitor therapeutic response at unprecedented resolution, inform us why specific patients or cancer types are resistant and distill from this knowledge novel biomarkers predictive of response to checkpoint inhibitors. In the long run, we also expect these insights to reveal novel treatment combinations that provide long-term therapeutic responses in refractory patients.

Single cell profiling

We are currently profiling tumor biopsies from patients receiving checkpoint immunotherapy in clinical trials of breast cancer, recurrent head-and-neck, cervical cancer, ovarium cancer, metastatic lung cancer, (colo)rectal cancer etc. Interestingly, in 2 other cancer types, i.e., renal cell carcinoma and hepatocellular carcinoma, checkpoint immunotherapy is combined with an anti-angiogenic tyrosine kinase inhibitor, allowing us to explore the interaction between anti-angiogenic and checkpoint immunotherapies.


Circulating tumor DNA as a diagnostic biomarker for cancer behavior

Plasma cell-free DNA (cfDNA) can be used to characterize and monitor tumors non-invasively, providing a real-time representation of the entire tumor. While the plasma cfDNA in healthy individuals mainly originates from leukocytes, a substantial part of this DNA in cancer patients derived from tumor cells (ctDNA, i.e. circulating tumor DNA). Such a blood-based test detecting ctDNA is important, as most patients do not undergo a tumor re-biopsy. Instead, only the initial diagnostic biopsy is available and the molecular phenotype of the tumor may have changed over time. Drawing on rapid developments in the field of non-invasive prenatal testing (NIPT), cfDNA-based tests have gained momentum in the oncology space, as a potential cancer-screening tool. In this context, we have recently demonstrated the utility of the plasma-derived cfDNA low-coverage whole-genome sequencing (LC-WGS) analysis using chromosomal instability as read out in the ovarian cancer setting (Vanderstichele et al., Clin Cancer Res. 2017). However, this approach is limited to the detection of tumors characterized by a certain degree of chromosomal instability and these will be missed when using this approach.

Copy numbers

In order to more generically detect ctDNA, we also explored the characterization of the tissue-of-origin of ctDNA by analyzing the patterns of the cell type-specific nucleosome positions in cfDNA using the same LC-WGS data. Fragmentation patterns of plasma-derived cfDNA are known to reflect nucleosome positions of cell types contributing to cfDNA. Since a substantial fraction of cfDNA from cancer patients originates from tumor cells, the nucleosome footprints in cfDNA are different between cancer patients and healthy controls.


As such, LC-WGS of cfDNA represents a single diagnostic test that generate 2 independent diagnostic read outs. However, these LC-WGS approaches need sufficient cfDNA input. In addition to the cfDNA-derived LC-WGS read outs, we have also developed a unique protocol to reliably assess the methylation status of low concentrations of heavily fragmented cfDNA by targeted bisulfite sequencing. cfDNA from cancer patients also differs in its DNA methylation pattern from that of healthy patients, with a focal hypermethylation at tumor suppressor genes and a global hypomethylation. In contrast with the DNA hypomethylation as a ubiquitous cancer mark, regions of DNA hypermethylation are more tumor-specific and therefore more amenable for tumor (sub-)typing. Preliminary data has demonstrated the potential of this targeted methylation-based cfDNA test as biomarker for detection of tumor DNA in plasma from patients with invasive ovarian cancer, hepatocellular carcinoma etc. 


Interestingly, our recent preliminary data has further demonstrated that the assessment of the different metrics (chromosomal instability, nucleosome footprinting and DNA methylation) act complementary in the ovarian cancer setting. Some tumors are missed using one approach, but can be picked up with the other approach, making combining the three metrics more reliable in detecting ctDNA than either approach alone.

We are currently optimizing these plasma-derived ctDNA methods and apply them in other cancer types as the fraction of ctDNA depends on the type of cancer because not all cancer types release an equal amount of tumor DNA in the bloodstream.


Our work is funded by: