Spotlight on Computational Genetics: Linking Data Generation and Target Discovery Research
I used to joke that when I started in genetics I moved different amounts of clear liquid from one place to another, and now, as a computational geneticist, I move text files from one place to another. But just as that was not just any clear liquid, these are not just any text files; they carry complex, valuable information that we hope will lead to therapeutic discoveries that impact patients.
I work on the computational genetics team at Variant Bio, where we use statistical and computational approaches to analyze genetic and related data (which we sometimes call multiomics), with the goal of providing novel biological insights.
There are different approaches to identifying new medicines for development. One common approach is target-based drug discovery, where researchers identify a molecular target (for example a gene or gene product) that could be modulated to modify a disease. There are many methods to identify these drug targets. However, researchers have observed that drug targets with human genetic evidence are more likely to be successful in clinical trials.
With this in mind, our team works to identify novel drug targets with human genetic evidence from the start of the process, and then to translate genetic data into insights to help guide target validation and early discovery efforts.
Right now, we are excited to be working on some novel population data where we have whole genome sequencing, health data, metabolomics, and RNA sequencing all within the same population. It is very exciting to be able to integrate multiple data types as we explore the biology of complex disease.
Our team sits right in the middle of the process, which is really fun — and also has its challenges! We get the opportunity to interact with a variety of scientists across Variant Bio on any given day. Data doesn’t just flow in one direction, so we find ourselves having to iterate and try to support different teams in order to make timely decisions — and this requires us to constantly tweak our priorities to try and best serve the company’s efforts.
For example, we support the partnerships team on some study design aspects, and if we see a measure is working very well in one study, we may recommend that it be included in another unrelated study. Another example: our biology team may have questions on subgroup analyses that could help them pick a study design. This ability to work across teams is invaluable and can drive better solutions.
We are also very fortunate to work with amazing partners around the globe who have a wealth of experience on the health factors of local populations, for example around gout in French Polynesia and liver fibrosis in Uganda. Our partners can provide guidance on things like treatment approaches and self-care for common conditions like gout so we can develop a responsive data analysis plan. These insights are critical for designing projects, for the analysis of data, and for downstream inference. Quality downstream inference, or reaching conclusions about the data, is the difference between having “just a lot of text files” and generating valuable hypotheses or supporting data-driven decisions.
Computational genetics and the greater mission at Variant Bio
Variant Bio aims to leverage the power of human genetic diversity to develop life-saving therapies, and the computational genetics team contributes to this overarching mission in two key capacities.
First, we analyze project data as efficiently as possible when it is generated so we can make discoveries and guide inference. If we do this right, we are able to maximize the potential of the data by getting the right analysis in the hands of the right scientists as quickly as possible.
Second, we innovate on methods so we can gain more insights and maximize the return on investment for each project. In the past decade, the field of genetics has included many large biobank projects with genetic and multiomics data that have really opened up geneticists’ ability to understand complex disease. These are incredibly important projects and have led to many method advancements for “big data.” However, many of these big data methods have to be adjusted for data sets like those we analyze at Variant Bio in order to maximize our potential to make new discoveries. This may include making sure methods can accommodate more related individuals, or that we use methods that appropriately account for a population’s genetic history.
While these large biobank projects have been transformative in the field, one well noted gap is the lack of diversity in these data resources. Variant Bio is deeply committed to working with diverse populations on research, and to generating genetic and multiomics data sets based on these collaborations. Our team finds this to be a really exciting scientific challenge that we get the opportunity to engage with on a daily basis.
