Reporter on the road: a roadmap for collaboration

24 January 2019
Talking to others about my time at VIB, there is one thing that always comes back: The open and collaborative atmosphere. I believe that this allowed me to succeed in getting my PhD, and to make useful connections that would guide me in my career path. But could it be proven in a scientific way? In biology, we often use graph theory to highlight the relationships between biological entities. Could we apply this to labs as well?

Materials & methods 
I made a list of all authors of the papers I co-authored during my time at VIB. Every lab was counted, and for each lab, I identified the number of co-authors. Labs now could serve asnodes in a network and their  relationships would be co-authorships.

Results
The left panel of Figure 1 shows what my collaborative network looks like. Node size in this graph is a function of the number of collaborators I had in that specific lab. Obviously, I have more collaborators in my PhD lab (B) than my MSc lab (A) and what would later become my postdoc lab (C). Because I only took my own papers into account, my PhD lab has the highest degree, a measure of connectedness.

However, some other labs also have darker blue colors and thicker lines running back to my PhD lab. Line thickness in this plot corresponds to the number of papers (that I co-authored) shared between two labs. As you can see in the middle panel, node 3 and 6, which are both darker blue than the rest and have thicker links with my PhD lab, are the SWITCH and Tompa labs, respectively. Indeed, during my PhD, we started collaborations with these two groups, which resulted in a few joint publications and ongoing projects.

Discussion
Besides giving us a pretty network visualizing scientific collaborations, does this type of analysis provide us with any relevant information? It would be useful to extend the analysis to labs, centers or even the entire institute. Which labs cluster together in collaborative networks and have the strongest connections? We could also figure out which labs connect two distant networks, and try to encourage these ties to promote integration of those networks (e.g., collaborative grant opportunities). Additionally, we can start using such collaborative networks as roadmaps to study not only the flow of information (i.e., shared papers), but also people.



The right panel indicates that the parameters we discussed above may even predict the flow of people. My own career path is in red: I started in the Verstrepen lab, went to the Van Den Bosch lab and ended up leaving VIB for a postdoc abroad. I mentored two bright MSc students, Emiel Michiels and Mathias De Decker. They completed their theses in the Van Den Bosch lab with Kevin Verstrepen as a co-promoter. Emiel joined the
SWITCH lab, our strongest collaborator, while Mathias stayed in the lab but started working for the group of Philip Van Damme. Young talent flows through the VIB ecosystem from one hub to the other. On the other hand, the Van Den Bosch and Tompa lab strengthened their collaborative efforts by recruiting new talent from outside the VIB pool: Donya Pakravan is now pursuing a PhD at VIB co-supervised by both the Tompa and Van Den Bosch labs.

This very limited case study shows that the connections identified via this methodology do carry important information and highlight opportunities to foster and strengthen the exchange of knowledge and talent within the VIB pool.
​​


Steven Boeynaems is a VIB alumnus who worked at the Kevin Verstrepen Lab and the Ludo Van Den Bosch Lab. Recently he traded Belgium for the Californian sun. At Stanford University he keeps pursuing his passion for science and science communication.

Instagram: @steven.boeynaems
Twitter: @BoeynaemsSteven