Inter-species differences of co-expression of neighboring genes in eukaryotic genomes.

BACKGROUND: There is increasing evidence that gene order within the eukaryotic genome is not random. In yeast and worm, adjacent or neighboring genes tend to be co-expressed. Clustering of co-expressed genes has been found in humans, worm and fruit flies. However, in mice and rats, an effect of chromosomal distance (CD) on co-expression has not been investigated yet. Also, no cross-species comparison has been made so far. We analyzed the effect of CD as well as normalized distance (ND) using expression data in six eukaryotic species: yeast, fruit fly, worm, rat, mouse and human.
RESULTS: We analyzed 24 sets of expression data from the six species. Highly co-expressed pairs were sorted into bins of equal sized intervals of CD, and a co-expression rate (CoER) in each bin was calculated. In all datasets, a higher CoER was obtained in a short CD range than a long distance range. These results show that across all studied species, there was a consistent effect of CD on co-expression. However, the results using the ND show more diversity. Intra- and inter-species comparisons of CoER reveal that there are significant differences in the co-expression rates of neighboring genes among the species. A pair-wise BLAST analysis finds 8-30 % of the highly co-expressed pairs are duplicated genes.
CONCLUSION: We confirmed that in the six eukaryotic species, there was a consistent tendency that neighboring genes are likely to be co-expressed. Results of pair-wised BLAST indicate a significant effect of non-duplicated pairs on co-expression. A comparison of CD and ND suggests the dominant effect of CD.

Background

As a consequence of DNA sequencing activities, whole-genome sequences for many microbial organisms as well as eukaryotic species are available in publicly accessible databases. DNA microarray technology makes it possible to simultaneously monitor expression patterns of thousand of genes. Expression profiles combined with whole-genome information, especially map information, enable us to investigate a relationship between co-expression of genes and a chromosomal distance (CD).

In the pioneering work in this field, Cohen et al. (2000) and Kruglyak and Tang (2000) independently showed that in yeast (Saccharomyces cerevisiae), adjacent pairs of genes show correlated expression [1,2]. In the nematode worm (Caenorhabditis elegans), a study of the relationship between physical distance and expression similarity found many co-expressed pairs of neighboring genes within a distance range of 20 kbp [3]. Clustering of co-expressed genes has been found in humans (Homo sapiens) [4], worm [5] and fruit flies (Drosophila melanogaster) [6,7]. However, in mice and rats, an effect of CD on co-expression has not been investigated thoroughly yet. Also, no cross-species comparison has been made so far. We analyzed the effect of distance using 24 expression datasets in six eukaryotic species: yeast, fruit fly, worm, rat (Rattus norvegicus), mouse (Mus musculus) and human and investigated inter-species differences.

Methods

Datasets

We used publicly available expression data in six eukaryotic species: yeast (two sets [8,9]), fruit fly (two sets [10-12]), worm (one set [13]), rat (three sets [14-16]), mouse (five sets [17-20]) and human (eleven sets [17], [21-28]) (for details see Table 1). Genes' chromosomal positions in yeast and worm were obtained from protein tables at GenBank , and those in fruit fly and those in mouse, rat and human were obtained from FlyBase and from Resourcerer at TIGR , respectively.

Table 1

Dataset information

species	label	genes	pairs	microarrays	weight
human	Su (a) [17]	5510	770014	85	0.0482
human	Armstrong [21]	5492	765803	72	0.0406
human	Pomeroy [22]	3107	239951	266	0.0470
human	Ramasawamy [23]	6971	1217536	280	0.2508
human	Yeoh [24]	5503	768242	254	0.1436
human	Houmard [25]	15681	6344163	24	0.1120
human	Diette [26]	24810	15248313	29	0.3253
human	Dyrskjot [27]	3159	245900	40	0.0072
human	Alizadeh (a) [28]	2932	160672	31	0.0037
human	Alizadeh (b) [28]	2701	187917	35	0.0048
human	Alizadeh (c) [28]	3696	341437	67	0.0168
mouse	Su (b) [17]	7651	1616309	90	0.2437
mouse	Schinke (a) [18]	9124	2296857	30	0.1154
mouse	Schinke (b) [18]	7649	1615129	54	0.1461
mouse	Neptune [19]	13150	4744714	24	0.1908
mouse	Scott [20]	9124	2296857	79	0.3040
rat	Faden [14]	5519	884975	79	0.5741
rat	Almon [15]	4227	529217	47	0.2043
rat	Almon [16]	4227	529217	51	0.2216
worm	Kim [13]	15368	19987830	553	1
fruit fly	Montalta-He and Egger [10,11]	5534	3078761	31	0.8171
fruit fly	Arbeitman [12]	1138	133501	160	0.1829
yeast	Eisen [8]	2478	232362	79	0.4315
yeast	Cho [9]	6123	1422380	17	0.5685

Label provides the reference information. The genes, pairs and microarrays columns show the total number of the genes, pairs and microarrays in each dataset, respectively. Weight denotes the dataset's weight in each species.

Calculation of co-expression rate

We investigated only the genes of which the chromosomal position was known. In each dataset, the genes were divided into nsubs ets, each of which consisted of genes/ORFs on the same chromosome (Here, ndenotes the number of the chromosomes in that species). The Pearson correlation coefficient, r, and chromosomal distance, CD, measured in base pairs (bp) were calculated for every possible pair in a subset. After this calculation was repeated for all subsets, all results were merged and the total number of the pairs, N, was counted. The N pairs were sorted according to their values of r (regardless of CD), and then the top 20 % of the N pairs were selected from the sorted data as highly correlated (HC) pairs. The pairs were sorted into bins of equal sized intervals of CD. Because the six species have different genome compactness, different widths were used: 20 kbp for the mammals (human, rat and mouse), 2.5 kbp for the worm and fruit fly data and 1 kbp for yeast. To investigate an effect of CD on co-expression, we defined a co-expression rate (CoER) in each bin:

When the CoERs of the six species were plotted in a graph, the first bin of the mammals was further divided so that the starting point of the curves were almost the same.

As control groups, we selected two more groups of 0.2 N pairs from the sorted data: the lowest 20 % of N, negatively correlated (NC) pairs, and the 20 % of N centered at the median, zero-correlation (ZC) pairs. The CoER was calculated for these groups, and the results were compared with the HC group.

Weighted average

Because there were large differences in the number of the gene-pairs and microarrays in the datasets, we sought some measure of the reliability of each dataset. To do so, we used the product of two factors which inform regarding the breadth of the sampling of the organism's transcriptome: the numbers of the total pairs and microarrays in each dataset. The averaged CoER in each species was calculated using weights derived from the products. We considered the dataset having the largest weight in each species the most reliable one. It was expected that the CoER would scatter around 0.2 if there was no distance (or any other) effect on co-expression. To verify this, we also calculated the standard deviation of the CoER using the weights.

Normalized distance

When the six spec ies were compared a normalized distance (ND) was also used. The physical distance was normalized by the grand average of CD of all gene pairs next to each other in each species. The overall average was calculated as follows. First, the total length of a chromosome was divided by the number of genes on it. This was repeated for all chromosomes in a species and by averaging all values, we obtained the overall average in that species.

Gene ontology category

We also investigated Gene Ontology (GO) categories in the HC, NC and ZC groups in CD ranges between 0 and 20 kbp and between 980 and 1000 kbp. This analysis was carried out only for the most reliable dataset of yeast (Cho et al. 1998) and human (Diette, PGA Human CD4+Lymphocytes, ).

Pair-wise protein BLAST

Because duplicated genes are likely to be co-expressed due to their common history, we investigated how many pairs in the HC group were duplicated ones. In each of the yeast, worm and mouse datasets, HC pairs were sorted according to the CD (in ascending order) and the top 10,000 were selected for a pair-wise, stand-alone protein BLAST analysis. Of the 10,000, the pairs in which the protein sequences of both genes were known were subject to analyze. We employed a criterion previously proposed for the identification of duplicated genes in vertebrates [3,4]. Gene pairs having expected values (E) of less than 0.2 were deemed to be duplicated. Protein sequences were obtained from ExPASy Molecular Biology Server for mouse, from Wormpep.109 for worm and from Saccharomyces Genome Database for yeast.

Statistical test

To investigate intra- and inter-species differences, a multiple comparison of CoER was carried out using the Ryan procedure [29] at the significance level of α = 0.01. The underlying concept of this procedure is the incremental adjustment of the significance thresholds. Except for the significance thresholds, each of the tests was made in exactly the same manner as for a single pair. First, the highest and lowest CoERs were compared at the significance level of 2α / n(n-1), where n was the number of samples to be compared (in the inter-species comparison, n = 6). If the extremes differed significantly, we tested each of them against the CoER next to the other extreme at the significance level of 2α / n(n-2). If we found a significant difference in the previous step, we continued to test the highest CoER vs. the third lowest, the lowest vs. the third highest and the second highest vs. the second lowest using 2α / n(n-3). These steps were repeated until 2α / n(n-5). Because different bin widths were used in yeast, worm and fruit fly, the CoER was recalculated for the range between 0 and 20 kbp when the six species were compared. This comparison method was also employed to investigate intra-species differences in the eleven human datasets.

Results and Discussion

Co-expression rate

In all 24 datasets the distribution of Pearson correlation coefficient appeared to be a bell-shaped curve centered at approximately zero. In what follows, we will describe the results of the HC pairs unless otherwise noted. A higher CoER was obtained in the first bin than a long CD range between 980 k – 1000 kbp in all datasets.

In Fig. 1, the CoERs obtained from the eleven human datasets are plotted as a function of CD: Fig. 1(A),1(B) and 1(C) show the results of the HC, ZC and NC pairs, respectively. The CoERs from the HC pairs decreased as the CD increased although there were large swings in CoER in a CD range above 100 kbp. These swings were larger in smaller data sets because of stochastic effects and undersampling of the transcriptome. In contrast, in the CD range below 100 kbp, the CoERs from the ZC and NC groups were relatively flat (compare Fig. 1(A) with 1(B) and 1(C)).

Figure 1

The co-expression rates (CoERs) in the eleven human datasets are plotted as a function of chromosomal distance (CD): the results from the highly-correlated (HC) pairs (A), the zero-correlation (ZC) pairs (B) and the negatively-correlated (NC) pairs (C). The reference information is as follows: 1) Armstrong [21], 2) Ramasawamy [23], 3) Pomeroy [22], 4) Su (a) [17], 5) Yeoh [24], 6) Houmard [25], 7) Diette [26], 8) Dyrskjot [27], 9) Alizadeh (a) [28], 10) Alizadeh (b) [28], 11) Alizadeh (c) [28]. The weighted average and standard deviation of the CoER over the eleven datasets: the results from the HC pairs (D), the ZC pairs (E) and the NC pairs (F).

The weighted average and standard deviation over the eleven curves in (A), (B) and (C) are shown in (D), (E) and (F), respectively. The average in (D) showed the same tendency of gradual decrease to 0.2, whereas the averages in (E) and (F) showed no distance effect. The weighted averages for all six species are given in Fig. 2. The phenomena shown in Fig. 1 were consistently observed in all species. These results strongly suggest that co-expression of neighboring genes is common in many eukaryotic species and also that the CD, especially a short CD below 10 kbp, is associated with increased frequency of co-expression.

Figure 2

A comparison of the six species using the weighted average of each species. (A) the highly correlated pairs, (B) the zero correlation pairs and (C) the negatively correlated pairs. Note that worm2 represents the CoERs of the worm dataset without the pairs in operons and duplicates.

GO category

The results of pair-wise analysis of GO category appeared to be different in the yeast and human datasets (Table 2). In yeast, only the HC pairs shared the same category and most of the pairs were not duplicates. Four pairs were found in a long CD range (980 – 1000 kbp). In the human data, more variety was obtained and pairs having the same category were found even in the NC group. A category, GO:5887 (integral to plasma membrane), was seen in the three groups while GO:6954 (inflammatory response) was seen in the HC and ZC. In the long CD range, eleven pairs were found. The results suggest the human genome involves more complicated functional relationships over a substantial CD.

Table 2

Gene Ontology categories shared by the two genes in pairs with a CD below 20 kbp

(A) highly correlated pairs
		category	description	# pairs
h	molecular functions	GO:0005132	interferon-alpha/beta receptor ligand	2
h		GO:0005194	cell adhesion	2
h		GO:0008089	anterograde axon cargo transport	4
h	biological process	GO:0006954	inflammatory response	6
h		GO:0006955	immune response	3
h		GO:0007165	signal transduction	5
h	cellular component	GO:0005634	nucleus	5
h		GO:0005886	plasma membrane	5
h		GO:0005887	integral to plasma membrane	2
y	molecular functions	GO:0003735	structural cons tituent of ribosome	12 (1)
y	biological process	GO:0006412	protein biosynthesis	13 (1)
y	cellular component	GO:0005737	cytoplasm	12 (2)
(B) zero-correlated pairs
h	molecular functions	GO:0005200	cytoskeletal structural protein	3
h		GO:0008009	chemokine	2
h		GO:0008417	fucosyltransferase	2
h		GO:0015464	acetylcholine receptor	2
h	biological process	GO:0006954	inflammatory response	3
h		GO:0007048	oncogenesis	4
h		GO:0007165	signal transduction	8
h	cellular component	GO:0005882	intermediate filament	3
h		GO:0005886	plasma membrane	4
h		GO:0005887	integral to plasma membrane	5
(C) negatively correlated pairs
h	molecular functions	GO:0005624	membrane fraction	3
h	biological process	GO:0006508	proteolysis and peptidolysis	2
h		GO:0007165	signal transduction	2
h	cellular component	GO:0005624	membrane fraction	3
h		GO:0005887	integral to plasma membrane	2

A 'h' and 'y' in the left column denotes that the pairs were from the human (Diette [26]) and yeast (Cho et al. [9]) dataset, respectively. In yeast, none pair shared the same category in the ZC and NC groups. A number in parentheses in the right column indicates the number of duplicated pairs.

Table 3 summarizes the BLAST results. In yeast, about 780 out of the 10,00 pairs were deemed to be duplicated pairs (E < 0.2). In the worm dataset, 8,370 pairs were available for the analysis and 2,658 (31.8 %) were regarded as duplicates. In the mouse datasets, out of about 3,500 analyzed pairs, 11.2 % were putative duplicates. However, in the three species, most pairs had expected values larger than 1. These results indicate that there are many non-duplicate pairs in the HC group, suggesting that the high CoERs in the HC group were due to not only effects of duplicated pairs but also non-duplicate effects. The distributions of the expected values in the five mouse datasets were almost the same (Table 3), suggesting that differences in microarrays were not significant.

Table 3

Distribution of expected values obtained in pair-wise protein BLAST

			Expected value
species	label	# of analyzed pairs	0 – 0.2	0.2 – 0.4	0.4 – 0.6	0.6 – 0.8	0.8 – 1	> 1
yeast	Eisen [8]	10000	800	515	402	316	369	7598
yeast	Cho [9]	9980	768	422	377	325	322	7766
worm	Kim [13]	8370	2658	267	206	182	210	4847
mouse	Su (b) [17]	3631	437	151	111	116	101	2715
mouse	Schinke (a) [18]	3554	340	149	131	107	99	2728
mouse	Schinke (b) [18]	3599	396	139	110	109	129	2716
mouse	Neptune [19]	3122	344	118	115	77	101	2367
mouse	Scott [20]	3659	408	117	110	104	111	2809

Intra- and inter-species comparisons

In Fig. 1(D), the standard deviations for the first three bins are relatively large. For example, the CoERs in the first bin in the eleven datasets were in a range between 0.24 and 0.38. To investigate intra-species differences in humans, a multiple comparison with the Ryan procedure was carried out (α = 0.01). Forty-nine out of the 55 possible combinations were not significantly different. In the five mouse datasets, 12,778 HC pairs out of approximately 20,000 used for the BLAST analysis were commonly seen in two or more datasets. The distributions of the expected values in the five mouse datasets were almost the same (Table 3). This suggests that there was no significant intra-species difference. Accordingly, the noise in the microarray data and the differences in microarray design appear to have minor influences on our results.

The CoERs in the weighted averages of the six species were compared. The results of the intra- and inter-species comparisons indicate that there are significant differences (p < 0.01) in the CoER in a short CD range (0 – 20 kbp) between any pair among worm, mammal (human, rat and mouse), fruit fly and yeast except two pairs of (worm and rat), and (mouse and fruit fly) (Table 4). Although the rat CoERs in the first three bins are almost the same as those of the worm (Fig. 2(A)), the results using the normalized distance (Fig. 3) strongly suggest that the multicellular organisms except worm show similar CoERs. The CoERs of worm and yeast were much larger than the others for a ND range between 0.3 and 1. In yeast, which is a unic ellular eukaryote with a compact genome, the organization of coordinated gene regulation is probably different from the other species with more dispersed genomes. As previously reported [3], the worm genome involved much more duplicated pairs than the yeast and mouse (Table 3). According to Blumenthal et al. [30], the worm genome involves at least 1,000 operons, which correspond about 15 % of all C. elegans genes. After excluding the pairs in the duplicates and operons, the worm CoERs are similar to those of the other multicellular organisms (see worm2 in Fig. 2(A)), indicating that the duplicates and operons are the main reason for the larger CoERs in the worm. However, when the ND was used, the worm curves (both with and without duplicates and operons) were similar to the yeast data.

Table 4

The results of a multiple comparison with the Ryan procedure

	worm	rat	human	mouse	fly	yeast
worm			*	*	*	*
rat	p = 0.703				*	*
human	p < 0.001	p = 0.094			*	*
mouse	p < 0.001	p = 0.017	p = 0.250			*
fly	p < 0.001	p < 0.001	p < 0.001	p = 0.077		*
yeast	p < 0.001	p < 0.001	p < 0.001	p < 0.001	p = 0.001

A * denotes significant difference (p < 0.01).

Figure 3

The co-expression rates shown in Figure 2(A) were re-plotted against the normalized distance. A value on the horizontal axis can be smaller than 1 because all possible pairs next to each other were involved in the calculation (i.e. we did not exclude a non-coding region, cetromere, etc from the calculation). Note that worm2 represents the CoERs of the worm dataset without the pairs in operons and duplicates.

A comparison of Figs. 2(A) and 3 provides some information on mechanisms behind the distance effect. If the physical distance has the dominant effect (chromatin remodeling is a possible cause), the CoERs in Fig. 2(A) should be similar across the species. On the other hand, if the effect of ND is major, Fig. 3 should show similar curves. The actual results seem to be the former and suggest that the CD plays the dominant role. In the multicellular organisms, the CoERs were higher than 0.2 up to about 50 kbp whereas the yeast curve was flat above 10 kbp. The mechanism behind this difference is currently not clear, but there are some clues. For example, several factors have been identified as controlling localized gene transcription [31]. These include the size of euchromatic chromosome territories and the spacing of chromatin "insulators" which provide impedance to non-specific enhancer activity upon neighboring genes [32]. Variation across species of these factors could explain our findings although this has not been systematically studied to date. Further analysis is required to advance our understanding of the mechanisms.

Conclusions

In this study, the effect of chromosomal as well as normalized distance on co-expression was analyzed using expression data from six eukaryotic species. We confirmed that in the six species, there was an effect of distance on CoER and a consistent tendency that neighboring genes are likely to be co-expressed. The results of intra- and inter-species comparisons of CoER show that there are significant differences in the co-expression rates of neighboring genes among yeast, worm and the other multicellular organisms. The pair-wise protein BLAST analysis indicates that effects of non-duplicates are not negligible. A comparison between the effects of chromosomal and normalized distance revealed that the physical distance has the dominant effect.

Authors' contributions

YF conceived the methodology, performed the statistical tests and drafted the manuscript. HI developed all software and performed the bioinformatical analysis. ISK guided the study and coordinated the project. All authors read and approved the final manuscript.