Towards human-computer synergetic analysis of large-scale biological data.

Area of Science:

• Bioinformatics and Computational Biology
• Human-Computer Interaction
• Information Visualization

Background:

• Massive, complex biological data (genomics to structural biology) presents analysis challenges.
• Current paradigm of automated algorithms followed by expert review limits exploratory analysis and hypothesis formulation.
• Integrating domain expertise during data analysis is crucial for poorly understood biological processes.

Purpose of the Study:

• To develop and evaluate a novel design approach, experiential computing, for analyzing large-scale biological data.
• To integrate domain expertise with computational methods for enhanced data exploration and hypothesis generation.
• To create user-friendly systems that facilitate human-computer synergy in biological data analysis.

Main Methods:

• Utilized and extended experiential computing, combining information visualization and human-computer interaction with algorithms.
• Emphasized direct user interaction, unified query/presentation spaces, external contextual information assimilation, and user-directed exploration.
• Developed two prototype web applications: XMAS for time-series transcriptional data and PSPACE for protein structure-function relationships.

Main Results:

• XMAS and PSPACE demonstrate the efficacy of the experiential computing paradigm for analyzing complex biological data.
• Systems facilitate human-computer synergy, integrating domain knowledge with algorithmic operations for large-scale analysis and visualization.
• Case studies showcase analysis of gene expression and marine organism responses (XMAS), and protein structure-function relationships (PSPACE).

Conclusions:

• The proposed framework seamlessly integrates visualization, algorithms, and cognitive expertise for sense-making, exploration, and discovery.
• Domain insights combined with algorithms facilitate knowledge discovery and evidence-based hypothesis formulation.
• User studies show XMAS and PSPACE offer better interpretability and lower cognitive load compared to comparable systems.