The gene indiced during the sporulation are mainly species-specific genes but orthologous groups of genes, which are not straighforward
distinguished, can be revealed by a systematic cross-talk analysis between expression data and orthology properties.
Indeed, of all the genes that showed a significant change in mRNA levels during sporulation (G1 = 518 genes for the S. cerevisiae
and G2 = 724 genes for the S. pombe), only 85 genes for S. cerevisiae and 98 genes for S. pombe exhibited sequence
conservation between the two yeasts. This moderate overlap between the two sets of genes sharing a similar fucntion
(~ 15 %) may be explained by both the stringency of the procedure to identify orthologous gene pairs and a rather
long evolutionary distance (at least 400 million years between S. cerevisiae and S. pombe). Since the choice of an
appropriate cut-off value is inevitably subjective, we developed a multi-step procedure to progressively assess the
validity of the approach. First, the genes with a firm orthologous relationship were compared in a 3-dimensional space.
This allowed the identification of similarities and differences in expression between the two organisms
(see
Figure 5). To go further in this analysis, the cross-species comparison can be extended
to the other genes ignored in the first step, using a "gene-centric" comparison (Lelandais et al., 2003).
Briefly, the rationale of this approach is to compare the immediate "expression neighbourhood" of a given inter-organism
gene pair (called the seed genes). As an illustration of this strategy (see below), we can chose as seed genes the
gene pair DAL5 (YJR152W, S. cerevisiae) and SPCC417.10 (S. pombe), whose expression is conserved between the two yeasts
(
Figure 5B, 2). They are both induced rapidly and transiently after transfer to sporulation
medium. Then, using a correlation cut-off of 0.7, 24 genes were found in the immediate neighbourhood of the S. cerevisiae
seed gene DAL5 and 48 genes in the neighbourhood of the S. pombe seed gene SPCC417.10 (Figure 7C and D).
Based on Gene Ontology (GO) keywords (Ashburner et al., 2000) (Zeeberg et al., 2003), two biological processes were found
in common between these genes: "Metabolism" (GO:0008152) and "Transport" (GO:0006810). In particular, the sub-category
"Nitrogen Metabolism" (GO:0006807) was found to be statistically significantly over-represented than what would be
expected by chance (the calculated p-values using FuncAssociate (Berriz et al., 2003) are around 0.05, the usual
significant level). This strongly supports the idea that these "organism-specific genes" are also involved in the
adaptation process to nitrogen starvation. Such a rationale approach can be applied to other gene motifs of the first
step of the sporulation process. It thus confirms the fact that a large proportion of "organism-specific genes" is
actually involved in this first step when compared with the subsequent steps. This is in agreement with the quantitative
assessment of the E-value found for the early genes (see main Text). This observation which is probably biologically
relevant, certainly deserves further analysis.
This strongly supports the idea that these "organism-specific genes" are also involved in the adaptation process
to nitrogen starvation. We believe that our approach can offer a complete description of the two groups of
genes involved in the first step for the induction of sporulation. The fact that many genes differ between
S. cerevisiae and S. pombe in this first step might be biologically significant.
