MediSyn: uncertainty-aware visualization of multiple biomedical datasets to support drug treatment selection

Chen He, Luana Micallef, ZiaurRehman Tanoli, Samuel Kaski, Tero Aittokallio, Giulio Jacucci

Introduction

In the era of precision medicine, the ability to select the most effective drug for a specific genetic mutation is paramount. However, the necessary biological data is often fragmented across multiple disparate databases, such as the Cancer Genome Interpreter (CGI) and Drug Target Commons (DTC). While technologies like Linked Data (RDF) facilitate machine-to-machine communication, they do little to alleviate the "cognitive load" placed on researchers who must manually synthesize heterogeneous information.

Furthermore, biomedical data is rarely perfect; it is frequently characterized by missing information, inconsistencies between sources, and varying levels of clinical credibility. Traditional tools often hide these nuances, potentially leading to suboptimal clinical decisions. To address this, we developed MediSyn, an uncertainty-aware visualization tool designed to help researchers navigate the "messy" reality of clinical and experimental data.

The MediSyn Framework

MediSyn utilizes a web-based, matrix-based layout where columns represent targets (mutations or wild-type genes) and rows represent drugs. This structure is inherently scalable and makes knowledge gaps—represented as empty cells—immediately obvious to the user.

The MediSyn matrix interface integrating drug-mutation effects and potency levels.

The core innovation of MediSyn lies in its overlaid layers. Rather than presenting datasets side-by-side, the system integrates them within the same visual space. Following established principles of visual encoding, the system uses Hue (green for responsive, red for resistant) to indicate clinical drug-mutation effects from CGI, and Length/Position (bar charts) to represent quantitative potency and evidence levels from DTC. Every data point includes provenance links to original publication sources (DOIs), allowing for immediate verification.

Data Integration and the Challenge of Heterogeneity

MediSyn synthesizes two major manually curated datasets to provide a comprehensive view of the drug-target landscape:

  • Cancer Genome Interpreter (CGI): Focuses on clinical biomarkers and drug responses across various tumor types.
  • Drug Target Commons (DTC): Provides quantitative bioactivity data (such as $IC_{50}$ or $K_d$ values) from experimental assays.
Standard tabular views of CGI and DTC data which MediSyn aims to synthesize.

By integrating 536 mutations, 258 drugs, and over 2,300 drug-target relations, MediSyn transforms complex tabular data into a visual map that highlights both consensus and conflict.

Visualizing Uncertainty and Conflict

Unlike systems that attempt to "clean" data by resolving conflicts behind the scenes, MediSyn treats inconsistency as a salient feature. For instance, a cell might display a red "Resistant" background from CGI overlaid with a "Highly Potent" black bar from DTC. This visual dissonance alerts researchers that a drug's efficacy may be context-dependent or debated in the literature.

Additionally, the inclusion of a "Wild-Type" column alongside mutations allows researchers to quickly identify potential toxicity. Drugs that show high potency against both the mutation and the healthy wild-type gene may pose a higher risk of adverse effects, a critical consideration in drug selection and repurposing.

Evaluation and Key Findings

A study conducted with six domain experts compared MediSyn against the standard practice of searching individual database websites. The results demonstrated significant improvements in both performance and user experience:

  • Efficiency: Researchers were over 3x faster at selecting appropriate drugs using MediSyn (averaging 1m 57s) compared to the baseline (6m 22s).
  • Usability: MediSyn received high marks for ease of use (7/7) and overall satisfaction (5.5/7).
  • Discovery: The tool facilitated "serendipitous discovery," such as identifying Axitinib as a potential treatment for the T315I mutation—a finding supported by clinical literature but difficult to synthesize from raw tables.

Conclusion and Future Work

MediSyn demonstrates that making data uncertainty visible can actually improve decision-making efficiency. While the current matrix layout is highly effective for two to three datasets, future work will explore summary views to prevent visual clutter as more sources are added. Additionally, experts have noted the need to integrate patient-specific context—such as age and treatment history—to further bridge the gap between experimental data and clinical application.

Summary updated

Abstract

Background Dispersed biomedical databases limit user exploration to generate structured knowledge. Linked Data unifies data structures and makes the dispersed data easy to search across resources, but it lacks supporting human cognition to achieve insights. In addition, potential errors in the data are difficult to detect in their free formats. Devising a visualization that synthesizes multiple sources in such a way that links between data sources are transparent, and uncertainties, such as data conflicts, are salient is challenging. Results To investigate the requirements and challenges of uncertainty-aware visualizations of linked data, we developed MediSyn, a system that synthesizes medical datasets to support drug treatment selection. It uses a matrix-based layout to visually link drugs, targets (e.g., mutations), and tumor types. Data uncertainties are salient in MediSyn; for example, (i) missing data are exposed in the matrix view of drug-target relations; (ii) inconsistencies between datasets are shown via overlaid layers; and (iii) data credibility is conveyed through links to data provenance. Conclusions Through the synthesis of two manually curated datasets, cancer treatment biomarkers and drug-target bioactivities, a use case shows how MediSyn effectively supports the discovery of drug-repurposing opportunities. A study with six domain experts indicated that MediSyn benefited the drug selection and data inconsistency discovery. Though linked publication sources supported user exploration for further information, the causes of inconsistencies were not easy to find. Additionally, MediSyn could embrace more patient data to increase its informativeness. We derive design implications from the findings.
BibTeX
@article{he2017medisyn,
  author = {Chen He and Luana Micallef and ZiaurRehman Tanoli and Samuel Kaski and Tero Aittokallio and Giulio Jacucci},
  title = {{MediSyn: uncertainty-aware visualization of multiple biomedical datasets to support drug treatment selection}},
  journal = {BMC Bioinformatics},
  year = {2017},
  doi = {10.1186/s12859-017-1785-7},
}
Figure