Cracking the code of undiagnosed diseases
An interview with VIB Bioinformatician – Dr.House - Zakaria Eddafir
June 17, 2024
350 million people worldwide live with undiagnosed diseases. Genome sequencing can provide answers for 40% of children and adults with undiagnosed diseases, but around 60% still don't receive a diagnosis. The Wilhelm Foundation is dedicated to finding new ways and collaborations to address diseases that currently have no solutions. That’s led them to organizing a hackathon, an event where participants from diverse disciplines use computational tools and techniques to solve biological problems within a set timeframe. At this Undiagnosed Hackathon event, 126 clinicians, bioinformaticians, molecular biologists, scientists, developers, AI specialists, and individuals from 28 countries formed multidisciplinary teams to tackle the most challenging diseases.
Out of 42 cases, 10 patients received a genetic diagnosis, and an additional nine patients received a potential diagnosis. Zakaria Eddafir, a bioinformatician and PhD student at VIB-UAntwerp Center for Molecular Neurology, took part in the event and was interviewed by two national newspapers about his contribution, comparing him to the famous TV show Doctor “House”. He now shares his experience with us.
Read the Dutch interviews in Het Nieuwsblad and De Gazet Van Antwerpen.
Hi Zakaria, how did you get involved with this event?
I’m part of Sarah Weckhuysen's group. She leads the epilepsy clinic, and her lab studies genetic epilepsies and encephalopathies. Because we deal with rare diseases, we often collaborate with other researchers. I participated in a similar hackathon a few months ago, where I learned a lot and made valuable connections. So, when this event came up, I knew I had to join. It was a big event with many renowned participants, including new AI startups.
What was it like working in such a diverse team?
As a bioinformatician, I'm usually far removed from patients, working behind a screen on technical details. At the hackathon, it was eye-opening to see the clinicians so closely connected to the patients. They are deeply invested in the outcomes. When a candidate gene is identified, their emotional involvement is evident. We worked in groups, with some internal competition that really drove everyone to do their best.
Can you describe a specific challenge you faced?
Recognizing genetic variants is one thing, but determining their cause is another. As a bioinformatician, it sometimes seems obvious that a certain snip or translocation is the cause, but clinicians actually provide crucial insights here. They can assess if the phenotype matches. It's really a collaborative effort, with different opinions and many discussions.
What about the cases you solved?
Our team solved three cases and found a candidate variant in another. One interesting case involved a mutation that’s found in 0.5% of neurodevelopmental disorders and microcephaly cases, which isn’t all that rare! We also found out that re-analyzing data can also yield the solution. We consider a case solved when the phenotype matches and the mutation is a well-known variant in the literature. Afterwards, every finding is confirmed with Sanger sequencing.
What’s next for these families?
Although identifying a variant is important, actual treatments for the children are still a long way off. There are some possibilities, though. When a variant is identified, a clinician might recommend them for a clinical trial. For example, some progress has been made in trials using antisense oligonucleotides (ASOs). Knowing the episignatures may also be used to recommend certain medications. However, for most patients, it ends with the identification. We hope that discovering mutations in multiple cases will help to make new diagnoses and drive toward personalized medicine. This gives families renewed hope to explore what’s possible.
How did you get into bioinformatics?
I studied biomedical sciences at KU Leuven. In my first master’s year, we had to do three internships, and I was encouraged to try different things. I didn’t take to lab work, but then I did an internship and my thesis in the Laboratory of Multi-omic Integrative Bioinformatics (LMIB) in Leuven. I also did an internship in Sweden at the Lundeberg lab, working on spatial transcriptomics, which led me to this field. Doing a pure bioinformatics PhD with my background wasn’t straightforward, as we lack extensive statistics and machine learning training. On the other hand, my knowledge of biology comes in handy a lot of the time, which a ‘pure’ bioinformatician might not have.
What does your PhD focus on?
I’m currently doing a PhD that focuses on developing an integrated multi-omics pipeline for long-read sequencing data. The goal of this project is to better understand the genetics of early-onset epilepsy, identify disease pathways, and reduce the diagnostic gap. It’s 90% bioinformatics and 10% lab work, which I actually learned to love!
What do you do when you’re not at work?
I enjoy powerlifting as a way to relax and challenge myself. I also love watching anime. My favorite is Vinland Saga.
VIB is organizing some hackathons later this year. Stay tuned to this page for updates.