Michael Correll

Error Bars Considered Harmful: Exploring Alternate Encodings for Mean and Error.

When making an inference or comparison with uncertain, noisy, or incomplete data, measurement error and confidence intervals can be as important for judgment as the actual mean values of different groups. These often misunderstood statistical quantities are frequently represented by bar charts with error bars. This paper investigates drawbacks with this standard encoding, and considers a set of alternatives designed to more effectively communicate the implications of mean and error data to a general audience, drawing from lessons learned from the use of visual statistics in the information visualization community. We present a series of crowd-sourced experiments that confirm that the encoding of mean and error significantly changes how viewers make decisions about uncertain data. Careful consideration of design tradeoffs in the visual presentation of data results in human reasoning that is more consistently aligned with statistical inferences. We suggest the use of gradient plots (which use transparency to encode uncertainty) and violin plots (which use width) as better alternatives for inferential tasks than bar charts with error bars.

Perception of Average Value in Multiclass Scatterplots

The visual system can make highly efficient aggregate judgements about a set of objects, with speed roughly independent of the number of objects considered. While there is a rich literature on these mechanisms and their ramifications for visual summarization tasks, this prior work rarely considers more complex tasks requiring multiple judgements over long periods of time, and has not considered certain critical aggregation types, such as the localization of the mean value of a set of points. In this paper, we explore these questions using a common visualization task as a case study: relative mean value judgements within multi-class scatterplots. We describe how the perception literature provides a set of expected constraints on the task, and evaluate these predictions with a large-scale perceptual study with crowd-sourced participants. Judgements are no harder when each set contains more points, redundant and conflicting encodings, as well as additional sets, do not strongly affect performance, and judgements are harder when using less salient encodings. These results have concrete ramifications for the design of scatterplots.

High Genetic Diversity and Adaptive Potential of Two Simian Hemorrhagic Fever Viruses in a Wild Primate Population

Key biological properties such as high genetic diversity and high evolutionary rate enhance the potential of certain RNA viruses to adapt and emerge. Identifying viruses with these properties in their natural hosts could dramatically improve disease forecasting and surveillance. Recently, we discovered two novel members of the viral family Arteriviridae: simian hemorrhagic fever virus (SHFV)-krc1 and SHFV-krc2, infecting a single wild red colobus (Procolobus rufomitratus tephrosceles) in Kibale National Park, Uganda. Nearly nothing is known about the biological properties of SHFVs in nature, although the SHFV type strain, SHFV-LVR, has caused devastating outbreaks of viral hemorrhagic fever in captive macaques. Here we detected SHFV-krc1 and SHFV-krc2 in 40% and 47% of 60 wild red colobus tested, respectively. We found viral loads in excess of 106–107 RNA copies per milliliter of blood plasma for each of these viruses. SHFV-krc1 and SHFV-krc2 also showed high genetic diversity at both the inter- and intra-host levels. Analyses of synonymous and non-synonymous nucleotide diversity across viral genomes revealed patterns suggestive of positive selection in SHFV open reading frames (ORF) 5 (SHFV-krc2 only) and 7 (SHFV-krc1 and SHFV-krc2). Thus, these viruses share several important properties with some of the most rapidly evolving, emergent RNA viruses.

Conditional CD8+ T Cell Escape during Acute Simian Immunodeficiency Virus Infection

ABSTRACT CD8+ T cell responses rapidly select viral variants during acute human immunodeficiency virus (HIV)/simian immunodeficiency virus (SIV) infection. We used pyrosequencing to examine variation within three SIV-derived epitopes (Gag386-394GW9, Nef103-111RM9, and Rev59-68SP10) targeted by immunodominant CD8+ T cell responses in acutely infected Mauritian cynomolgus macaques. In animals recognizing all three epitopes, variation within Rev59-68SP10 was associated with delayed accumulation of variants in Gag386-394GW9 but had no effect on variation within Nef103-111RM9. This demonstrates that the entire T cell repertoire, rather than a single T cell population, influences the timing of immune escape, thereby providing the first example of conditional CD8+ T cell escape in HIV/SIV infection.

Task-driven evaluation of aggregation in time series visualization

Many visualization tasks require the viewer to make judgments about aggregate properties of data. Recent work has shown that viewers can perform such tasks effectively, for example to efficiently compare the maximums or means over ranges of data. However, this work also shows that such effectiveness depends on the designs of the displays. In this paper, we explore this relationship between aggregation task and visualization design to provide guidance on matching tasks with designs. We combine prior results from perceptual science and graphical perception to suggest a set of design variables that influence performance on various aggregate comparison tasks. We describe how choices in these variables can lead to designs that are matched to particular tasks. We use these variables to assess a set of eight different designs, predicting how they will support a set of six aggregate time series comparison tasks. A crowd-sourced evaluation confirms these predictions. These results not only provide evidence for how the specific visualizations support various tasks, but also suggest using the identified design variables as a tool for designing visualizations well suited for various types of tasks.

Two Novel Simian Arteriviruses in Captive and Wild Baboons (Papio spp.)

ABSTRACT Since the 1960s, simian hemorrhagic fever virus (SHFV; Nidovirales, Arteriviridae) has caused highly fatal outbreaks of viral hemorrhagic fever in captive Asian macaque colonies. However, the source(s) of these outbreaks and the natural reservoir(s) of this virus remain obscure. Here we report the identification of two novel, highly divergent simian arteriviruses related to SHFV, Mikumi yellow baboon virus 1 (MYBV-1) and Southwest baboon virus 1 (SWBV-1), in wild and captive baboons, respectively, and demonstrate the recent transmission of SWBV-1 among captive baboons. These findings extend our knowledge of the genetic and geographic diversity of the simian arteriviruses, identify baboons as a natural host of these viruses, and provide further evidence that baboons may have played a role in previous outbreaks of simian hemorrhagic fever in macaques, as has long been suspected. This knowledge should aid in the prevention of disease outbreaks in captive macaques and supports the growing body of evidence that suggests that simian arterivirus infections are common in Old World monkeys of many different species throughout Africa. IMPORTANCE Historically, the emergence of primate viruses both in humans and in other primate species has caused devastating outbreaks of disease. One strategy for preventing the emergence of novel primate pathogens is to identify microbes with the potential for cross-species transmission in their natural state within reservoir species from which they might emerge. Here, we detail the discovery and characterization of two related simian members of the Arteriviridae family that have a history of disease emergence and host switching. Our results expand the phylogenetic and geographic range of the simian arteriviruses and define baboons as a natural host for these viruses. Our findings also identify a potential threat to captive macaque colonies by showing that simian arteriviruses are actively circulating in captive baboons.

Exploring collections of tagged text for literary scholarship

Modern literary scholars must combine access to vast collections of text with the traditional close analysis of their field. In this paper, we discuss the design and development of tools to support this work. Based on analysis of the needs of literary scholars, we constructed a suite of visualization tools for the analysis of large collections of tagged text (i.e. text where one or more words have been annotated as belonging to a specific category). These tools unite the aspects of the scholars’ work: large scale overview tools help to identify corpus‐wide statistical patterns while fine scale analysis tools assist in finding specific details that support these observations. We designed visual tools that support and integrate these levels of analysis. The result is the first tool suite that can support the multilevel text analysis performed by scholars, combining standard visual elements with novel methods for selecting individual texts and identifying represenative passages in them.

Quantity estimation in visualizations of tagged text

A valuable task in text visualization is to have viewers make judgments about text that has been annotated (either by hand or by some algorithm such as text clustering or entity extraction). In this work we look at the ability of viewers to make judgments about the relative quantities of tags in annotated text (specifically text tagged with one of a set of qualitatively distinct colors), and examine design choices that can improve performance at extracting statistical information from these texts. We find that viewers can efficiently and accurately estimate the proportions of tag levels over a range of situations; however accuracy can be improved through color choice and area adjustments.

Visualizing virus population variability from next generation sequencing data

Advances in genomic sequencing techniques allow for larger scale generation and usage of sequence data. While these techniques afford new types of analysis, they also generate new concerns with regards to data quality and data scale. We present a tool designed to assist in the exploration of the genetic variability of the population of viruses at multiple time points and in multiple individuals, a task that necessitates considering large amounts of sequence data and the quality issues inherent in obtaining such data in a practical manner. Our design affords the examination of the amount of variability and mutation at each position in the genome for many populations of viruses. Our design contains novel visualization techniques that support this specific class of analysis while addressing the issues of data aggregation, confidence visualization, and interaction support that arise when making use of large amounts of sequence data with variable uncertainty. These techniques generalize to a wide class of visualization problems where confidence is not known a priori, and aggregation in multiple directions is necessary.

LayerCake: a tool for the visual comparison of viral deep sequencing data

MOTIVATION The advent of next-generation sequencing (NGS) has created unprecedented opportunities to examine viral populations within individual hosts, among infected individuals and over time. Comparing sequence variability across viral genomes allows for the construction of complex population structures, the analysis of which can yield powerful biological insights. However, the simultaneous display of sequence variation, coverage depth and quality scores across thousands of bases presents a unique visualization challenge that has not been fully met by current NGS analysis tools. RESULTS Here, we present LayerCake, a self-contained visualization tool that allows for the rapid analysis of variation in viral NGS data. LayerCake enables the user to simultaneously visualize variations in multiple viral populations across entire genomes within a highly customizable framework, drawing attention to pertinent and interesting patterns of variation. We have successfully deployed LayerCake to assist with a variety of different genomics datasets. AVAILABILITY AND IMPLEMENTATION Program downloads and detailed instructions are available at http://graphics.cs.wisc.edu/WP/layercake under a modified MIT license. LayerCake is a cross-platform tool written in the Processing framework for Java. CONTACT mcorrell@cs.wisc.edu.