Comparative analysis of multiple genome-scale data sets

Abstract

The ongoing analyses of published genome-scale data sets is evidence that different approaches are required to completely mine this data. We report the use of novel tools for both visualization and data set comparison to analyze yeast gene-expression (cell cycle and exit from stationary phase/G(0)) and protein-interaction studies. This analysis led to new insights about each data set. For example, G(1)-regulated genes are not co-regulated during exit from stationary phase, indicating that the cells are not synchronized. The tight clustering of other genes during exit from stationary-phase data set further indicates the physiological responses during G(0) exit are separable from cell-cycle events. Comparison of the two data sets showed that ribosomal-protein genes cluster tightly during exit from stationary phase, but are found in three significantly different clusters in the cell-cycle data set. Two protein-interaction data sets were also compared with the gene-expression data. Visual analysis of the complete data sets showed no clear correlation between co-expression of genes and protein interactions, in contrast to published reports examining subsets of the protein-interaction data. Neither two-hybrid study identified a large number of interactions between ribosomal proteins, consistent with recent structural data, indicating that for both data sets, the identification of false-positive interactions may be lower than previously thought.

Figure 1

α-Factor–arrest data set (18 time points) ordinated and visualized in VxInsight. (A) Cell-cycle gene expression after α-factor arrest and the dendogram indicating similarities of gene expression as presented by Spellman et al. (Reprinted, with permission, from Spellman et al. 1998.) (B) Three-dimensional topography in which mountains are formed over clusters of genes. The height of the mountain corresponds to the number of genes beneath it. Typical expression profiles for genes in each mountain are provided. G₁, S, and M: Genes in these clusters are induced during the G₁, S, or M phase of the cell cycle, respectively. (C) Ordination of genes (dots) that underlie the topography with links (blue lines with yellow arrows at each end) showing strong similarities (Pearson's R > 0.887) that exist between genes in different clusters.

Figure 2

VxInsight-generated ordination of exit from stationary-phase data set. Examples of gene expression within each hill or cluster are shown. Along the x-axis of insert graphs are time points (0, 15, 30, 45, and 60 min) after re-feeding. The y-axis of insert graphs indicates the fold-increase or decrease from time equals; 0, which is an average of four to five replicates for each time point. Numbers in the insert graphs indicate the maximum value of the y-axis, which indicates relative expression values obtained using GeneSpring (Silicon Genetics; see Methods). Data were generated as described (Methods).

Figure 3

Location of G₁-regulated genes in two different gene-expression data sets. (A) Dots represent selected G₁-regulated genes in α-factor–arrest cell-cycle data (Spellman et al. 1998). (B) Location of the same genes in the ordination of stationary-phase exit data.

Figure 4

Location of ribosomal protein genes (RPS genes) in two gene-expression data sets. (A) Location of RPS genes in exit from stationary phase data. Fifty-three of 59 RPS genes are localized in the upper middle cluster. (B) Localization of the same RPS genes in cell-cycle data set. Arrows indicate three major groups of RPS genes.

Figure 5

Protein-protein interaction maps as a function of the cell-cycle gene-expression topography. Lines are drawn between genes encoding interacting proteins. (A) Schwikowski's complete data set. (B) Ito's full data set. (C) Protein-protein interactions reported from both data sets. (D) Genes encoding interacting proteins common to both data sets. In A and B, genes encoding proteins involved in interactions are indicated by yellow pyramids.

Figure 6

Interactions among proteins encoded by G₁-regulated genes from the cell-cycle data set. (A) Topographical presentation of G₁-regulated gene cluster with connections between genes showing strong similarities (R > 0.887) of expression between genes. (B) Genes encoding interacting proteins from Ito's full data set. (C) Genes encoding interacting proteins reported from Schwikowski's data set. (D) Protein interactions in common to the two data sets. Connections between genes in B–D indicate interactions occurring between proteins encoded by the specific genes.

Figure 7

Protein-protein interactions between Nup116p and other proteins as a function of gene expression. (A) Ito's full data set: cell-cycle expression topography. (B) Schwikowski's full data set: cell-cycle topography. (C) Ito's full data set: exit from stationary phase topography. (D) Diagram of Nup116p interactions in the nuclear pore from the Munich Information Center for Protein Sequences (http://vms.gsf.de/htbin/search_code/YMR047C). (Reprinted, with permission, from E. Hurt, BZH; Universitaet Heidelberg.)