Feature-expression heat maps – A new visual method to explore complex associations between two variable sets

• Published in: Journal of biomedical informatics 2015-02, Vol.53, p.156-161
• Main Authors: Haarman, Bartholomeus C.M. (Benno), Riemersma-Van der Lek, Rixt F., Nolen, Willem A., Mendes, R., Drexhage, Hemmo A., Burger, Huibert
• Format: Article
• Language: English
• Subjects: Algorithms; Associations; Cluster Analysis; Clusters; Color; Computational Biology - methods; Computer Graphics; Data Collection; Data Display; Data visualization; Exploration; False Positive Reactions; Gene Expression Profiling; Genetic Association Studies; Genetics; Genotype; Graph; Graphical representations; Heat map; Hot Temperature; Humans; Mathematical models; Method; Models, Genetic; Monocytes - cytology; Pattern recognition; Phenotype; Regression Analysis
• ISSN: 1532-0464; 1532-0480
• DOI: 10.1016/j.jbi.2014.10.003

[Display omitted] •Feature-expression heat maps provide insight into complex associations.•It utilizes effect ordered data display on two variable sets.•Effect size (color) and statistical significance (radius) are depicted in circles.•Its applications are found in complex biological systems. Existing methods such as correlation plots and cluster heat maps are insufficient in the visual exploration of multiple associations between genetics and phenotype, which is of importance to achieve a better understanding of the pathophysiology of psychiatric and other illnesses. The implementation of a combined presentation of effect size and statistical significance in a graphical method, added to the ordering of the variables based on the effect-ordered data display principle was deemed useful by the authors to facilitate in the process of recognizing meaningful patterns in these associations. The requirements, analyses and graphical presentation of the feature-expression heat map are described. The graphs display associations of two sets of ordered variables where a one-way direction is assumed. The associations are depicted as circles representing a combination of effect size (color) and statistical significance (radius). An example dataset is presented and relation to other methods, limitations, areas of application and possible future enhancements are discussed. The feature-expression heat map is a useful graphical instrument to explore associations in complex biological systems where one-way direction is assumed, such as genotype-phenotype pathophysiological models.