In this three-part feature series, we meet members of our postdoc and PhD community - to learn who they are, what roles they play in our research and what inspires them to pursue a career in science.
In 2016, Fanny Krebs attended a summer school for PhD students in San Sebastian, Spain. She had an organic chemistry background but was increasingly fascinated by molecular modelling, and signed up for a workshop on drug design that Vincent Zoete, with Antoine Daina from his SIB lab, were chairing. Fanny had a multitude of questions for him during the event and their exchanges led him to suggest she apply for a postdoctoral position he was opening in Lausanne. With a resilience forged in the face of substantial health concerns following a lab accident during her master’s and a learned awe for her field, Fanny is among the young scientists contributing to computer-aided molecular engineering research of the Zoete Lab in Lausanne, for the benefit of cancer patients.
What led you to choose a career in computational modelling?
I had always been interested in the development of drug molecules as my studies in organic chemistry led me to work on the development of antibiotics. At that stage though, I saw computational tools as the “dark side of the Force”, and it would be no exaggeration to say that I really did not have a clue about computers! But I was nevertheless intrigued by them since they offered the means to predict the movement of proteins and their binding.
In the course of my master’s, I carried out a practical experiment, which through no fault of my own, led to a major explosion in the lab. I was lucky to survive but suffered serious health consequences, which involved several operations and ensuing limitations on my ability to work in chemistry contexts with solvent fumes and strong lighting. A year into my PhD, it became clear that I had to reorient.
Around the same time, I was introduced to Roland H. Stote of the IGBMC in Strasbourg. Through his mentoring, I discovered the universe of molecular modelling, how much you can do with the help of computers if you know how to use them, and I realised that this was the future. All “hard” sciences - biology, chemistry, experimental - conducted in wet labs are hugely time- and resource consuming. In our field, with just a cluster of computers, we can often save an incredible amount of time in merely identifying possible approaches to problems.
The Zoete Lab teams work in a highly advanced field, can you describe that?
The two laboratories directed by Vincent (at SIB with Olivier Michielin, and at the Department of oncology/Ludwig Lausanne Branch) gather experts in molecular modelling and bioinformatics. The main projects relate to the development of computer-aided algorithms, programs and databases for protein engineering and drug design, with applications in oncology. The SIB group, the oldest, has made major contributions to the community through its development of great webtools like SwissSimilarity.ch, SwissADME.ch or SwissTargetPrediction.ch. The UNIL and Ludwig Lausanne Branch group is dedicated to immunology and precision oncology projects. Here, we work on the development of algorithms, to predict the binding of molecules to proteins with applications notably in immuno-informatics, or the potential effect of uncharacterized mutations based on structural and physics-based approaches.
In such a computational context, is there room left for the human touch?
Definitely. Computational power (still!) needs the worth of human thinking and our ability to appraise identified approaches. We bring our experience, and our intuition – which you could define as a non-verbal expression of our long experience.
Sometimes I notice mutations on a protein that are in a particular zone and find myself thinking: “No way… something’s going to happen here, this mutation is bound to have an impact on the protein activity and lead us to a clinical outcome.” Driven by such hunches, we can then conduct tests to see if, in silico, we do indeed observe a potential effect. And we’re often right.
What skills would you say it takes to undertake a career in science?
You need curiosity, creativity, passion and perseverance, – being able to accept, and not get discouraged, when things don’t immediately work out are vital to a career in research. And a great sense of communication, not only with colleagues in your own lab, but further afield too; building your network is so crucial. A morning chat with someone at the coffee machine, or the courage to approach a speaker at a conference, all are entry points into a vast knowledge pool you should absolutely immerse yourself in.
What is it that you like about being part of Zoete Lab?
The projects we work on are fascinating, and all the team members are at the forefront of their domain. I even feel that they are not aware of just how good they are, which means they remain extremely accessible and collaborative. If certain projects do not work out, we exchange and find solutions together. Vincent himself is an extraordinary leader and scientist, and whilst managing two labs, he still keeps a foot in research, which is a mighty balancing act.
What do you most enjoy about your work as a postdoc?
I love the fact that I am constantly learning. New tools come out every day and it is essential to keep abreast of those. Also, I get to see the utility of what we do, through the direct effect of our work on others’ lives. It’s important to me to be a link in the chain that helps a patient, or helps others to build on the tools we have created. It is the exact reason I chose a career in the public health sector, as opposed to one in financial sector where our profiles are highly sought after.
I’m also conscious of a continual sense of marvel. Every day, I get to observe how our bodies and immune systems are battling against this system of proliferation that is cancer. Our proteins – our checkpoints – are relentlessly regulating this throughout our lives. Our bodies are in fact constantly fighting cancer. When a cancer actually manages to develop, it is down to a disruption of this precision system. I guess that makes us detectives of sorts, processing all the mutation “clues” that we pick up on that can, or not, allow us to go on to predict their impact.
What do you find most challenging about your work?
Time, or rather the lack of it. Every Monday, I make myself a checklist for the week. Come Friday, there is always more I wish I could have done!
How do you see your field evolving in the coming years?
I think more and more collaboration will be forthcoming and necessary, as we capitalise on the mind-boggling power of machine learning and AI. I hope that in all this we won’t lose sight of the centrality of the human dimension that I mentioned. That’s where those communication skills come in: through our shared experiences, we get to keep a critical mind. Because there are always things that we see that others do not, and vice versa. I’m all about technology, but we need to keep our feet on the ground.
Looking to the future, what is your personal objective?
To work with the team on improving the tools we are currently developing, their capacity of prediction for the mutations, and to accelerate the interpretation of data for the patients’ benefit.
And finally, what advice might you give to your 18 year old self?
Have the courage to pursue what interests you, even if it seems a steep curve. Always trust yourself on that draw you feel, and don’t be scared of failure. It is not because you do not pass an exam, or lose a year that you should give up. Take a step back, or even a complete break, and just allow yourself the time to see things differently. Challenges are there to build us into who we are. I had some hefty setbacks early on but it is the choices I made in the face of those that have led me to the career and the fulfilment I have today.
On being a postdoc in Zoete Lab...
A large part of my work as a postdoc in Vincent’s lab is devoted to the weekly Tumor Molecular Board (TMB), organised by the Lausanne and Geneva University Hospitals. It brings together oncologists of the region’s university hospitals and private clinics who need us to analyse (genetic) mutations they have identified in patients’ biopsies. This allows the TMB to evaluate if there are any associated therapeutic solutions. These mutations might already be documented, but we don’t know if we can use a particular treatment in their presence as the approach has yet to be tested. Or, the mutations have yet to be described in the literature which means we don’t know their impact on the patient’s cancer cells as a whole; sometimes it can even be both.
The analysis we perform studies mutations at the level of their protein structures to predict whether they will have a harmful effect and if this might constitute a good rationale for a treatment. We draw on numerous international databases to do this.
Another role I have is to develop our own webtool that references known mutations, structures and sequences of identified onco-driver genes. A first online version, Swiss-PO.ch, is already available to the scientific community. We’re currently working on an update to include new genes, a scoring function to predict the potential damaging effect of a mutation and all the mutation analyses we have ever come across via the TMB. To give you an idea of what that represents: in the last five years, we have carried out over 900 analyses! With our overarching perspective on our own and other databases, we even occasionally spot mutations overlooked at the bioinformatics stage that could be a target or that could potentially activate a protein, which in turn could become a target.
This interview was conducted in French. All images, including @Twitter, are courtesy of Fanny Krebs and Zoete Lab.