Visualization is an interdisciplinary field related to data science and cognition, concerned with visual representations of data. Visualization research is often conducted in collaboration with researchers from other disciplines, for example, climate and environmental sciences, urban planning, and medicine. One such application area is genomics data, which is increasingly being used within healthcare for diagnosis and treatment selection. However, genomics data is often large and error-prone and requires interpretation to be of use to the treating physician. These challenges can be addressed with visualization. 

This thesis maps out the challenges and requirements of genomics data in clinical contexts from a visualization perspective. It contains an analysis of the tasks users performed and what data features were important to them. Several visualization environments for genomics data were designed and evaluated, based on these analyses. In general, the emphasis of the designs was on position and navigation, along with data aggregation and comparison of different data types. This was especially important for the work on copy number variant review, where several data sources were combined in a glyph. Therefore, genomics visualizations often benefit from combining multiple views, zooming, and different visual representations. 

Several of the challenges that were encountered in developing the genomics visualization environments generalize to other sequence data visualizations, such as timelines. Therefore, we studied how the sequence axis should be designed for these visualizations in general. In one study, graphs with straight layouts were compared to circular layouts. Another study investigated whether the x-axis could be stretched and compressed to reduce whitespace. The findings indicated that compressing axes can increase item visibility while the viewer maintains an understanding of the data distribution.  

A prominent recurring theme in the thesis is the utilization of position and the possibility of modifying the graph axes. Which axis type is most appropriate depends on the data, the task, and the user. In particular, it depends on how important the position along the axis is in comparison with all other data features. For genomics data, the position is typically not the most important feature. Therefore, axis manipulations can be both justified and appropriate.