Visualization of three-way comparisons of omics data - PubMed (original) (raw)

Comparative Study

Visualization of three-way comparisons of omics data

Richard Baran et al. BMC Bioinformatics. 2007.

Abstract

Background: Density plot visualizations (also referred to as heat maps or color maps) are widely used in different fields including large-scale omics studies in biological sciences. However, the current color-codings limit the visualizations to single datasets or pairwise comparisons.

Results: We propose a color-coding approach for the representation of three-way comparisons. The approach is based on the HSB (hue, saturation, brightness) color model. The three compared values are assigned specific hue values from the circular hue range (e.g. red, green, and blue). The hue value representing the three-way comparison is calculated according to the distribution of three compared values. If two of the values are identical and one is different, the resulting hue is set to the characteristic hue of the differing value. If all three compared values are different, the resulting hue is selected from a color gradient running between the hues of the two most distant values (as measured by the absolute value of their difference) according to the relative position of the third value between the two. The saturation of the color representing the three-way comparison reflects the amplitude (or extent) of the numerical difference between the two most distant values according to a scale of interest. The brightness is set to a maximum value by default but can be used to encode additional information about the three-way comparison.

Conclusion: We propose a novel color-coding approach for intuitive visualization of three-way comparisons of omics data.

PubMed Disclaimer

Figures

Figure 1

Figure 1

Examples of color-codings for three-way comparisons. Color representations for three-way comparisons of selected values a, b, and c calculated using the proposed procedure are shown in the column labeled HSB-based. Colors acquired by substituting values of a, b, and c directly for red, green, and blue or cyan, magenta, and yellow (black = 0) are shown in columns labeled RGB or CMYK, respectively. The legend is drawn as a hexagon instead of a circle for convenience. Horizontal lines separate groups of values with similar distributions.

Figure 2

Figure 2

Metabolite profiles for the three-way comparison. Mouse liver extract metabolite profiles acquired by CE-TOFMS two hours after intraperitoneal injection with (a) vehicle (Control), (b) diethylmaleate (DEM), a non-protein thiol-depleting chemical, or (c) buthionine sulfoximine (BSO), an inhibitor of γ-glutamylcysteine synthase. The plotted datasets are averages of five normalized replicate datasets for cation measurements originating from our previous work [2]. The averaged datasets are visualized as density plots. For all plots, numbered ovals (annotation labels) indicate the expected locations of peaks of a set of known chemical compounds and are used for identification of metabolites on the density plots [2,3].

Figure 3

Figure 3

Three-way comparison of metabolite profiles. (a) Absolute × relative three-way comparison of metabolite profiles shown in Figure 2. Averages of replicate datasets (n = 5) were used for the three-way comparison. The resulting dataset was filtered using F-ratio (one-way ANOVA) to select only statistically significant differences as described in the main text. (b) The Control dataset (Figure 2a) was overlaid on the three-way comparison result shown in panel (a) via the brightness value. Darkening of the colored spots indicates the size of the corresponding peaks in the Control dataset. Gray spots show peaks which do not significantly differ among the datasets. For both plots, numbered ovals (annotation labels) indicate the expected locations of peaks of a set of known chemical compounds and are used for identification of metabolites on the density plots [2,3].

Figure 4

Figure 4

Candidate differences. Overlaid extracted ion electropherograms for the most significant differences from the three-way comparison results shown in Figure 3. Each panel represents data in the form of signal intensity (number of ions) over time for a specific mass interval (1 Da bin). The vertical dashed line indicates the position of the most significant difference according to the three-way comparison results. When present within panels, numbers correspond to the annotation labels in Figures 2 and 3.

Similar articles

Cited by

References

    1. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998;95:14863–14868. doi: 10.1073/pnas.95.25.14863. - DOI - PMC - PubMed
    1. Soga T, Baran R, Suematsu M, Ueno Y, Ikeda S, Sakurakawa T, Kakazu Y, Ishikawa T, Robert M, Nishioka T, Tomita M. Differential metabolomics reveals ophthalmic acid as an oxidative stress biomarker indicating hepatic glutathione consumption. J Biol Chem. 2006;281:16768–16776. doi: 10.1074/jbc.M601876200. - DOI - PubMed
    1. Baran R, Kochi H, Saito N, Suematsu M, Soga T, Nishioka T, Robert M, Tomita M. MathDAMP: a package for differential analysis of metabolite profiles. BMC Bioinformatics. 2006;7:530. doi: 10.1186/1471-2105-7-530. - DOI - PMC - PubMed
    1. Hedenfalk I, Duggan D, Chen Y, Radmacher M, Bittner M, Simon R, Meltzer P, Gusterson B, Esteller M, Kallioniemi OP, Wilfond B, Borg A, Trent J. Gene-expression profiles in hereditary breast cancer. N Engl J Med. 2001;344:539–548. doi: 10.1056/NEJM200102223440801. - DOI - PubMed
    1. Kyng KJ, May A, Kolvraa S, Bohr VA. Gene expression profiling in Werner syndrome closely resembles that of normal aging. Proc Natl Acad Sci USA. 2003;100:12259–12264. doi: 10.1073/pnas.2130723100. - DOI - PMC - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources