Therefore, it was long assumed that eukaryotic genes were randomly distributed across the genome due to the high rate of chromosome rearrangements.
Half a year after that Velculescu et al. (1997) published a research in which they had integrated SAGE data with the now available genome map.
Caron et al. (2001) made a human transcriptome map of 12 different tissues (cancer cells) and concluded that genes are not randomly distributed across the chromosomes.
They identified 27 clusters of genes with very high expression levels and called them RIDGEs.
Clusters of genes in close proximity and high transcription levels can easily been generated by tandem duplicates.
These duplicates can either became a functional part of the pathway of their parent gene, or (because they are no longer favored by natural selection) gain deleterious mutations and turn into pseudogenes.
Versteeg et al. (2003) tried, with a better human genome map and better SAGE taqs, to determine the characteristics of RIDGEs more specific.
[5] LINE repeats are junk DNA which contains a cleavage site of endonuclease (TTTTA).
Versteeg et al. also concluded that, contrary to Lerchers analysis, the transcription levels of many genes in RIDGEs (for example a cluster on chromosome 9) can vary strongly between different tissues.
They found seven universally clustered pathways: glycolysis, aminoacyl-tRNA biosynthesis, ATP synthase, DNA polymerase, hexachlorocyclohexane degradation, cyanoamino acid metabolism, and photosynthesis (ATP synthesis in non plant species).
[8] This study also showed that within clusters the expression levels of on average 15 genes was much the same across the many experimental conditions which were used.
These similarities were so striking that the authors reasoned that the genes in the clusters are not individually regulated by their personal promoter but that changes in the chromatin structure were involved.
[9] Many genes which are grouped into clusters show the same expression profiles in human invasive ductal breast carcinomas.
Contrary to previous discussed reports Johnidis et al. (2005) have discovered that (at least some) genes within clusters are not co-regulated.
It functions in negative selection of thymocytes, which responds to the organisms own epitopes, by medullary cells.
A very effective way to test this would be by insert synthetic genes into RIDGEs, antiridges and/or random places in the genome and determine their expression.
As an insertion construct they used a fluorescing GFP gene driven by the ubiquitously expressed human phosphoglycerate kinase (PGK) promoter.
[11] They investigated if these differences in expressions were due to genes in the direct neighborhood of the constructs or by the domain as a whole.
They also checked if the construct was expressed at similar levels as neighboring genes, and if that tight co-expression was present solely within RIDGEs.
Previous observations and the research of Gierman et al. proved that the activity of a domain has great impact on the expression of the genes located in it.
The genes of the research of Johnidis et al. were dependent of the present of the aire transcription factor.
Abderrahim et al. (1994) had shown that all the genes of the major histocompatibility complex were clustered on the 6p21 chromosome.
Yamashita et al. (2004) showed that genes related to specific functions in organs tend to cluster.
Six liver related domains contained genes for xenobiotic, lipid and alcohol metabolism.
Five colon-related domains had genes for apoptosis, cell proliferation, ion transporter and mucin production.
Cohen et al. found that of a pair of co-expressed genes only one promoter has an Upstream Activating Sequence (UAS) associated with that expression pattern.
In comparison with this model Gilbert et al. (2004) showed that RIDGEs are mostly present in open chromatin structures.
How eukaryotic gene regulation, and associated chromatin changes, precisely works is still very unclear and there is no consensus about it.
Singer et al. proposed that genes came in close proximity by random recombination of genome segments.
It has been observed in yeast and worms that essential genes tend to cluster in regions with a small replication rate.