The TGMI Team – Márton Münz

Marton Munz TGMI TeamEvery month, one member of the TGMI team will tell us why they are so committed to the vision of the TGMI, and share a bit more about their work and interests. This week we hear from Márton Münz, bioinformatician and software developer in the TGMI team.


What has been the main focus of your work to date?

I have been working as a bioinformatician and software developer for the Genetic Susceptibility Team at the Institute of Cancer Research over the past 5 years. My main areas of focus have been developing a standardised system for variant annotation, improving quality control of next-generation sequencing (NGS) data and implementing automated NGS data processing pipelines. Curiously, during the past 10 years my interests have gradually shifted from theory to practice – it must somehow be related to my personality development. After my PhD in structural bioinformatics and systems biology (which I spent studying conformational flexibility of proteins using molecular dynamics simulations), I ultimately realised that I wanted to work on questions that were more directly related to clinical application. The Mainstreaming Cancer Genetics programme and the TGMI provided a great opportunity for me to get involved with translational research projects of clinical relevance. The software tools we have developed at ICR are now routinely used in the clinic to test for cancer predisposition genes in breast and ovarian cancer patients.


What are you most excited about in genetic medicine?

Data scientists will be the new heroes of the field.

As a computational person, I have a good appetite for data crunching, so I am most excited about the thought that in the near future there will be a lot more data to crunch. I think genetic medicine is currently at a tipping point. It has already been shaken up by a huge data explosion (see the recent rapid growth of genomics databases), but in the next 5-10 years this will further increase by multiple orders of magnitude, as high-throughput sequencing is getting increasingly cheaper and widely used. This tsunami of data is going to transform genetic medicine beyond recognition. Data scientists will be the new heroes of the field. The fact that sequencing and genetic testing are soon becoming routine will not only allow us to look at the personal genetic information of patients, but also to interpret those in the context of a very large number of sequenced exomes or genomes. This will require a robust and highly reliable informatics infrastructure.


What are you most concerned about in genetic medicine?

In order to leverage the wealth of sequencing data, an important step is to integrate datasets from a wide variety of sources. For instance, aggregating large variant databases seems to be key to facilitate variant interpretation (and to avoid variant overinterpretation). However, data integration is currently painfully challenging due to the heterogeneity of the data. Lots of variations exist between sequencing platforms, data processing tools (read mappers, variant callers, annotation tools), the transcript databases used for variant annotation, and between the ways variants are represented. These discrepancies not only make data integration challenging – and benchmarking new tools difficult – but also lead to non-reproducible and inaccurate clinical test results. Standardisation of the reporting of genetic variations could help a lot to minimise these errors.


Why did you get involved in TGMI?

I think the main idea of the TGMI is simple: before we can reliably share data between each other, we must first agree on a common “language” to use. As simple as it sounds, standardisation turns out to be a really difficult problem. Standards have to both be unambiguous and need to capture all the inherent complexity of variant reporting. This looked like an exciting problem to me and I immediately felt I wanted to be involved.


What is the most important thing that you would like the TGMI to achieve?

To recommend reasonable standards for the clinical genetics community to adhere to when analysing and reporting genetic information, and to offer software tools that implement these standards.


If you had a magic wand (i.e. unlimited people/resources) what would you do to make genetic medicine work?

I would set up a large interdisciplinary group of geneticists, bioinformaticians and machine learning experts to focus on variant interpretation. In particular, the group would develop machine learning tools to predict disease risk from a large set of common low-penetrance genetic variants. It has been demonstrated by several studies that the collective effect of a large number of common variants can explain a major part of the heritability in many complex diseases, even if these variants only have very small effects individually. Currently, however, common variants are not typically included in clinical disease risk prediction. I think the complex relationships between combinations of common variants and disease risk could be best captured by complex machine learning (e.g. deep learning) models.


Do you have a favourite gene? If so – what and why?

I promised myself: if I miraculously recover, I would forget quantum mechanics and study genes instead

COL4A5 – which encodes the Alpha5 subunit of Type IV Collagen – is my favourite. In some sense, it is a rather boring gene with a long repetitive region, but it definitely helped in my career choice. At the age of 21, I was diagnosed with Alport Syndrome and developed end stage kidney failure. I was in the middle of my physics Masters at the time and was very much into quantum mechanics. If I had not received that diagnosis, I would probably have become a theoretical physicist. From that point, however, I only wanted to understand my disease. I admired my doctors in the dialysis centre who understood what was happening inside my body. When I read a bit on the topic, and found out that my life was likely falling apart because of a mutation in my COL4A5 gene , I was fascinated. I promised myself: if I miraculously recover, I would forget quantum mechanics and study genes instead.


What is a surprising fact that few people know about you?

Before jumping into my scientific career, I used to work as a science journalist. I worked for daily and weekly newspapers and also for a radio and a TV programme.


If you had a chance to experience a completely different career for a week, what job would you try?

I would like to be a novelist – although I know one week is not enough to write a novel. Perhaps I would try to write a short story.