Combined analysis reveals a core set of cycling genes

Abstract

Background: Global transcript levels throughout the cell cycle have been characterized using microarrays in several species. Early analysis of these experiments focused on individual species. More recently, a number of studies have concluded that a surprisingly small number of genes conserved in two or more species are periodically transcribed in these species. Combining and comparing data from multiple species is challenging because of noise in expression data, the different synchronization and scoring methods used, and the need to determine an accurate set of homologs.

Results: To solve these problems, we developed and applied a new algorithm to analyze expression data from multiple species simultaneously. Unlike previous studies, we find that more than 20% of cycling genes in budding yeast have cycling homologs in fission yeast and 5% to 7% of cycling genes in each of four species have cycling homologs in all other species. These conserved cycling genes display much stronger cell cycle characteristics in several complementary high throughput datasets. Essentiality analysis for yeast and human genes confirms these findings. Motif analysis indicates conservation in the corresponding regulatory mechanisms. Gene Ontology analysis and analysis of the genes in the conserved sets sheds light on the evolution of specific subfunctions within the cell cycle.

Conclusion: Our results indicate that the conservation in cyclic expression patterns is much greater than was previously thought. These genes are highly enriched for most cell cycle categories, and a large percentage of them are essential, supporting our claim that cross-species analysis can identify the core set of cycling genes.

Figure 1

Method overview. (a) Genes (nodes in the graph) are connected to other genes based on sequence similarity. Species identity is indicated by shape of nodes. Genes are also connected to a 'score node', which represents cycling expression score. Information is propagated along the edges until convergence. Genes are assigned a posterior score and a cut-off is applied to select the top genes for each species. (b) The subgraph containing the selected genes is further analyzed by identifying multidomain homology cliques. Examples of identified cliques of conserved genes are presented in panels c to f. (c) Cyclins. Fission yeast Cig2 promotes the onset of S phase [45]. Human Ccna2 is part of the G2 checkpoint [46]. (d) Cdc6/Cdc18 is a conserved and essential component of pre-replication complexes (pre-RCs). Orc1 is the largest subunit of the origin recognition complex (ORC), which binds specifically to replication origins and triggers the assembly of pre-RCs [47]. (e) TOG related proteins, a family of microtubule-associated proteins (MAPs). Proteins in this group localize to the plus-end tips of microtubules and are essential for spindle pole organization. Alp14 is a component of the Mad2-dependent spindle checkpoint cascade sharing redundant functions with Dis1. Mutants with both genes knocked out are nonviable [48]. (f,g) Microtubule component clique and expression profiles for fission yeast Nda3 in eight experiments [4-6]. Nda3, a known cell division gene [49], obtains a high cycling score but is not one of the 600 top cycling fission genes based on expression analysis. Using our method, its score is correctly elevated because its sequence similarity to high scoring genes.

Figure 2

Analysis of cycling genes using complementary high throughput datasets. (a) Number of interactions between cycling genes and nine cell cycle transcription factors. (b) Average expression level of sets of budding yeast genes in stationary phase (data from Gasch and coworkers [50]). (c) Expression levels of human genes in normal tissues, using data presented by Shyamsundar and colleagues [51] (also see Additional data file 1 [Supporting Figure 4]). Genes in the conserved set have lower expression levels for most nonproliferating normal tissues when compared with the full list and the list presented by Whitfield and coworkers [1]. For 26 out of 36 normal tissues this difference is significant with a P value < 0.05. (d) Arabidopsis cells in developmental arrest experiments [52]. Flower cells in the mutants stop growing after stage 11, whereas cells in the stem grow normally. Again, the conserved set is expressed at lower levels in developmental arrest (P = 0.027 at stages 11 and 12; P = 0.003 at stages 13 and 14). (e) Expression data from studying sexual differentiation and mating in fission yeast [53].

Figure 3

Conservation of cycling genes. (a) Percentage of conserved cycling genes in the four species. (b) Enrichment of cell cycle related Gene Ontology GO terms between all cycling genes and the CCC3 set in budding yeast, fission yeast, and humans. (c) Yeast protein-protein interactions [18]. We counted the number of interactions within a random set of 80 cycling yeast genes. In all, 1,000 sets were sampled. The histogram on the left plots the number of interactions observed for these sets. X represents internal interactions with the CCC3 set, which has significantly more internal interactions.

Figure 4

The importance of the core cycling genes. (a) Percentage of essential genes in different sets of budding yeast genes [29]. Although 18% of budding yeast genes are essential, only 15% of cycling genes are essential. Our analysis resolves this apparent contradiction by showing that the conserved cycling genes lists contain a much higher percentage of essential genes (35% and 46% for CCC3 and CCC4). Sequence alone cannot account for this high percentage (27%), indicating the importance of the combined analysis. (b) Similar analysis for the human lists using data from RNA interference knockdown experiments [30].