Mapping the Universe - Mapping Scientific Frontiers: The Quest for Knowledge Visualization - page 82

Fig. 2.35 Caenorhabditis elegans gene expression terrain map created by VxInsight, showing
three-dimensional representation of 44 gene mountains derived from 553 microarray hybridiza-
tions and consisting of 17,661 genes (representing 98.6 % of the genes present on the DNA
microarrays) (Reprinted from Kim et al. 2001 )

correlation coefficient between every pair of genes. For each gene, the similarity
between it and the 20 genes with the strongest positive correlations were used to
assign that gene to an x-y coordinate in a two-dimensional scatter plot with the use
of force-directed placement. Each gene is placed to other genes that are similar in
gene expression. Figure 2.35 shows a terrain map of Caenorhabditis elegans gene
expressions.

2.4.4 A Map of Influenza Virus Protein Sequences

In May 2009, as H1N1 was rapidly spreading across many countries, there was a
rich body of knowledge about influenza pandemics in the literature. Figure 2.36
shows a similarity map of 114,996 influenza virus protein sequences. Each dot is
an individual influenza virus protein sequence. Two sequences are connected if they
are similar in terms of protein structure. Structural similarity is one way to organize
protein sequences. There could be other ways, for example, based on similarities
of biological properties. Once again, multiple perspectives can be applicable. The
question is what would be the best combination of information provided by various
views to solve problems at hand.