“The idea for this project had been simmering for quite some time already. My research group had studied the association between genomic variants and gene expression at the population level. As you probably know, the genome of one person displays millions of tiny differences from the genome of another person. These differences are called ‘genomic variants’. When located in the protein-coding parts of the DNA, genomic variants may be associated with phenotypic variation such as ABO blood type or genetic disorders such as Huntington disease. However, around 97% of genomic variants are located in the non-coding parts of our DNA. It is much more difficult to interpret these variants. The non-coding DNA is the realm of the enhancers, promoters, and silencers that regulate gene expression, but it also contains large segments that do not appear to have any function. Hence, it is hard to tell whether non-coding variants have functional effects. My lab has developed a technology to assess this at the level of complete genomes.”
 
“At an Oncode meeting, I spontaneously got talking to Michiel Vermeulen of Radboud University. His group is specialized in studying the mechanistic aspects of how transcription factors bind to regulatory elements of the genome. These are molecular biological studies in biochemical assays. Our approaches nicely complement each other, so we started to collaborate. This quickly went off and we managed to publish our results in a nice journal in 2019.” 

“Cells acquire tens of thousands of mutations on the journey from healthy cell to pre-tumour cell to tumour cell. When located in coding DNA, it is relatively easy to establish the functional effects of these mutations. It is much more difficult to interpret the non-coding mutations. Mapping the association of these DNA variants with gene expression may greatly increase our insight into their role in the disease process. Enter the third group on our running train: the group of Lude Franke at UMCG. Lude is a geneticist and a renowned expert in bioinformatics. His group had already helped us with the analysis of our previous data. The three of us realized that our complementary tools and expertise would enable us to assess the functional effects of non-coding variants in genomes of cancer cells. Next, I talked with Sarah Derks of VUmc at an Oncode dinner. She is a physician scientist that works on gastroesophageal cancer. This type of cancer often cannot be explained by coding variants, so her group was a natural addition to our team, and they became the fourth partner.”

In 2021, Oncode launched a call for Oncode Accelerator Projects. This call was not associated with funding, but rather aimed to bring together scientists across different disciplines and fields of expertise. Oncode organized a brainstorm session, where Oncode Investigators pitched their research ideas. At this event, we explicitly searched for natural partners that would further complement the expertise of our team. The groups of Emile Voest, Jeroen de Ridder, and Wilbert Zwart joined us, each bringing unique complementary expertise. After consulting Hugo Snippert and others, we wrote an exciting project proposal, aiming to investigate non-coding variants in tumour cell genomes of patients with gastric or intestinal cancer. In the short term, we would like to unravel which non-coding variants in these cancer genomes have regulatory effects and establish the nature of these effects. Ultimately, this may provide crucial information that clinicians can use to select the most promising therapy for individual patients. Our dream is to develop an algorithm that interprets tumour sequencing data and flags points of attention for clinicians. This may be feasible within ten years or so.” 
 
“We managed to attract external funding through the Antoni van Leeuwenhoek Foundation, and the project kicked off in Fall 2021. The first datasets have been generated, and together with the Zwart Group – who are experts in genomics analysis of tumour samples – we are ramping up these efforts. Our unique technology produces immensely large datasets, with millions of measurements. The groups of De Ridder and Franke lead the bioinformatics analysis of the data, and the Vermeulen Group will perform the mechanistic interpretation. The Voest and Derks Groups bring in their view on what matters to patients. In addition, they are experts in using organoids and clinical materials to verify the results of the computational models. So, our groups really form a natural match rather than the type of forced collaboration that you often see in consortia. 

In my opinion, it is absurd if cancer researchers compete. This is not what public research money is meant for. Therefore, I actively contact peer researchers if I discover that they are working on overlapping topics, so that we can coordinate our efforts, collaborate, and exchange expertise. My lab also runs an Open Science Blog, where we try to inform people of ongoing work. We hope that this contributes to a more productive, collaborative, open kind of science. Collaboration is a strong means to realize goals and it can be a lot of fun. However, collaborations do call for investments of energy and time, as well as good communication and clear agreements. So, I try to choose my research partners wisely and search for those that are truly complementary as well as committed. Being part of Oncode forges a bond between different labs across the country, promoting this type of commitment.”