ABySS-Explorer: Visualizing Genome Sequence Assemblies

Abstract

Fig. 1. ABySS-Explorer employs a novel graph representation enabling biologists to examine the global structure of a genome sequence assembly.Abstract-One bottleneck in large-scale genome sequencing projects is reconstructing the full genome sequence from the short subsequences produced by current technologies. The final stages of the genome assembly process inevitably require manual inspection of data inconsistencies and could be greatly aided by visualization. This paper presents our design decisions in translating key data features identified through discussions with analysts into a concise visual encoding. Current visualization tools in this domain focus on local sequence errors making high-level inspection of the assembly difficult if not impossible. We present a novel interactive graph display, ABySS-Explorer, that emphasizes the global assembly structure while also integrating salient data features such as sequence length. Our tool replaces manual and in some cases pen-and-paper based analysis tasks, and we discuss how user feedback was incorporated into iterative design refinements. Finally, we touch on applications of this representation not initially considered in our design phase, suggesting the generality of this encoding for DNA sequence data.

An extended evaluation of the readability of tapered, animated, and textured directed-edge representations in node-link graphs

“…Our follow-up study, furthermore, showed C b providing no significant improvement over C [14]. Nielsen et al [20] note that they were inspired by our initial-study results. For their ABySSExplorer they use tapering to depict the orientation of "contigs," i.e., genome assemblies consisting of long contiguous sequences.…”

Design Study Methodology: Reflections from the Trenches and the Stacks

“…Abstraction should happen very early in the discovery stage and should frequently be validated by checking back with the expert to ensure correctness. One design study where we put significant effort into articulating the abstractions clearly and concisely was MizBee [45]; ABySS-Explorer is an excellent example from other authors [56].…”

“…A task abstraction will either be a good or a bad reflection of the actual domain, while a data abstraction proposed by the researcher will be appropriate or inappropriate for the specific problem at hand. An interesting example is the ABySS-Explorer project [56] where the computational algorithms that assemble a genome from sequence data rely on a graph with unique sequence strands represented as nodes and overlapping strands as edges. The visualization researchers observed, however, that the experts often used a dual representation (swapping nodes and edges) when sketching their ideas on paper, and found that building their visualization tool around this swapped representation allowed the domain scientists to reason about the output of the algorithms very effectively.…”

“…Design study papers are explicitly welcomed at several visualization venues as a way to explore the choices made when applying visualization techniques to a particular application area [55], and many exemplary design studies now exist [17,34,35,56,94]. A careful reading of these papers reveals multiple steps in the process of conducting a design study, including analyzing the problem, abstracting data and tasks, designing and implementing a visualization solution, evaluating the solution with real users, and writing up the findings.…”

Travel-Time Maps: Linear Cartograms with Fixed Vertex Locations

“…Lately, more specifically, the use of circular arcs for edges has received attention (e.g., [7]), in particular the creation of Lombardi drawings [6]. The use of regular sinusoid curves to indicate relative length was recently introduced by Nielsen et al [14] for the visualization of connectivity graphs in genome sequencing. Lastly we note the topic of map labeling, where a suitable position of each label must be found among a set of candidate positions (see [20] for a survey).…”