Segerdell Thesis Proposal Defense

How can an interactive visual analytics toolhelp biomedical scientists investigategenotype-phenotype relationships?

Thesis Proposal

Defense

Erik Segerdell, DMICE MS Student

April 3, 2015

I will present my proposal for a usability study concerned with genotype-phenotype investigation

1

2

3

4

Background

Specific Aims

Project Plan & Outcome

Timeline

The relationship between phenotype and genotype is fundamental to many areas of the life sciences

[ bbc.co.uk ]

Genotype Phenotype

Codes for

We seek better bioinformatics tools to explore and gain insights from genotype and phenotype data

Genotype Phenotype

Codes for

[ bbc.co.uk ]

Model organism databases contain comprehensive information about genotypes and phenotypes

[ zfin.org ]

Grids are one technique for visualizing and interacting with phenotypic data

[ monarchinitiative.org ]

Visual analytics tools should help foster hypothesis generation

Visualization tools need to support “explanatory” analysis and synthesis

Interactivi

ty

Updated interactions across views, e.g. across overlaid networks.

Facile filtering and color-coding.

Simplification of networks through zooming, filtering, and node

expansion or contraction.

Inference of biologically significant sets of relationships based on the

identification of indirect interactions or recurrent ones.

Content Weighted edges in the graph corresponding to biological traits.

Overlays of protein-protein interactions, disease associations, and

pathways.

Indications of functional distinctions between proteins and other bio-

molecules.

Based on Mirel 2009

Two initiatives involving OHSU research groups are exploring genotype-phenotype relationships

[ Bottomly et al. 2013 ]

[ monarchinitiative.org ]

Segerdell thesis proposal

1

2

3

4

Background

Specific Aims

Project Plan & Outcome

Timeline

The proposed work will answer my research question

How can an interactive visual analytics tool

help biomedical scientists investigate

relationships between genotype and

phenotype?

?

How can tools help foster hypothesis generation?

To answer the research question, I will pursue three specific aims

1. Determine design of a visualization

environment.

2. Create a functional, interactive prototype.

3. Evaluate the prototype.

Segerdell thesis proposal

1

2

3

4

Background

Specific Aims

Project Plan & Outcome

Timeline

The first aim is to determine design of avisual analytics prototype

Objectives?

Stakeholders and users?

Use cases?

Features to test?

Evaluation criteria?

Low-fidelity mockupswill form basis for functional ones.

I will develop storyboards to mock up entire beginning-to-end scenarios

[ www.sciencebuddies.org ]

The second aim is to create a functional, interactive prototype of the analytics tool

High-fidelity prototype(s)will be based on themockups.

Constructed with D3.

This will test the technicalfeasibility of theInteractive features.

[ github.com/mbostock/d3 ]

The third aim is to evaluate the usability of the prototypes

Evaluation will comprise three proven “discount” usability inspection methods

Think-Aloud Protocol

Semi-Structured Interview

Heuristic Evaluation

Evaluation will comprise three proven “discount” usability inspection methods

InterviewHeuristics Think-Aloud

A heuristic inspection will be will be conducted based on Nielsen’s guidelines

[ Nielsen 1995; planbozchi24.wordpress.com ]

A heuristic inspection will be will be conducted based on Nielsen’s guidelines

[ Nielsen 1995; planbozchi24.wordpress.com ]

A heuristic inspection will be will be conducted based on Nielsen’s guidelines

[ Nielsen 1995; planbozchi24.wordpress.com ]

The think-aloud protocol will get instant feedback from users about a small number of UI features

Groups of users (~5 or 6 total)• Students• Researchers• Clinicians

Users talk aloud as they carry out tasks based on concrete use cases.

• Describe actions• Explain/interpret actions

[ www.userjourneys.com ]

A semi-structured interview will take a general measure of the prototypes’ usability

[ Brooke 1996 ]

System Usability Scale (SUS)

Followed by debriefing interview.

A semi-structured interview will take a general measure of the prototypes’ usability

Ten items total.

Computed score is between 1-100.

[ Brooke 1996 ]

A semi-structured interview will take a general measure of the prototypes’ usability

[ Brooke 1996 ]

The work is expected to have the following outcome

• Qualitative analysis of a visual analytics tool containing interactive features for exploring genotypes and phenotypes.

• A case study of how someone used the tool, what worked, and what did not.

• Recommendations relevant to future interface development and quantitative evaluations.

✓

Segerdell thesis proposal

1

2

3

4

Background

Specific Aims

Project Plan & Outcome

Timeline

Timeline

SA1

SA2

SA3

Acknowledgments

Thesis Advisory Committee

• Beth Wilmot, PhD• Melissa Haendel, PhD• Harry Hochheiser, PhD• Michelle Hribar, PhD

Colleagues

• Monarch Initiative• Bioinformatics Shared

Resource, KCI• Ontology

Development Group, OHSU Library

• DMICE Faculty & Staff