Over 90% of the human and other vertebrate genomes don't have any known functional role [1, 2], and as such were considered until recently as "Junk DNA" or genomic "dark matter". A large fraction of the "non-functional" DNA originated from mobile elements  such as Alu, which comprise about 10% of the nucleotides of the human genome  with over one million inserted copies. This abundance is somewhat of a surprise, since Alu is a non-autonomous retroelement, i.e. it doesn't encode proteins that assist its mobilization, and therefore it needs to rely on the cell's machinery for its duplication in the genome . Currently Alu is believed to be a parasite of the mobilization machinery of long interspersed elements (LINE) .
A typical Alu is a dimer, comprised of a central A-rich region which is flanked by two similar sequence elements of about 130 bp (left and right arms). During evolution the Alu elements have spread in the human genome in several bursts at different times, which allow their classification into several families and subfamilies, on the basis of their insertions, deletions and mutations. The major families are: old AluJ, whose members are AluJb and AluJo (both spread in the genome 80–100 mya); the middle, AluS family, which includes AluSx, AluSg, AluSp AluSc and AluSq (35 –50 mya), and the youngest Alu family, AluY (25 mya) [5, 6].
One of the biggest mysteries associated with Alu concerns the non-random manner in which these elements are distributed throughout the human genome [1, 7]. The distribution of Alu was found to be highly correlated with local GC content and with the density of genes and with intron density . Furthermore, the density of Alu is higher in intragenic (12.5% of the nucleotides) than in intergenic regions (9.6% of the nucleotides) , which is surprising if indeed Alu elements are non-functional bits of DNA. In addition, Alu elements are negatively selected in imprinted regions of primate genomes . Transposable elements are underrepresented in the region surrounding the TSS of genes . Another analysis revealed that almost 20% of the genes contain transposable elements in the 3' UTR . Significant differences were observed also when Alu density was compared among full length gene sequences of different biological process classes (this study  was limited to chromosomes 21 and 22). Genes that were related to metabolism, transport and signalling were mostly Alu rich, whereas genes that had poor Alu content belonged mostly to structural proteins and to information processing and storage pathways . It was also shown recently that housekeeping genes contain more Alu than tissue specific genes .
This non-random of distribution of Alu, in particular the high concentration near and within genes, might be the result of positive feedback; selective pressure drives the organism to use these elements for some new functions, and the functions acquired by Alu generate advantage for phenotypes that have high Alu concentration near genes. Indeed, the possible functional roles of Alu in genomic organization and gene expression has received increasing attention during the last two decades (see [14, 15] for reviews). One of the more fascinating possible consequences of integration of mobile elements near genes is acquisition of a transcriptional regulatory role by the Alu element, as a carrier of different TF binding sites. The Alu element harbours BSs of several transcription factors that were shown to bind to Alu; most of these TFs were nuclear factors, hormones, calcium nuclear factors, as well as other TFs [16–22]. The idea of Alu acting as a carrier of cis regulatory elements was suggested by Britten  and by Oei at el.  who found YY1 and SP1 BSs on young Alu family members. These experimental results imply that Alu elements have influence on the regulation of expression levels of many genes that have Alu in their promoter regions; however, such a wide ranged regulatory impact has not yet been tested experimentally.
The main aim of our study was to find new putative BSs that reside on Alu. Very recently, several groups have performed similar analysis for all transposable element classes; however, the set of PSSMs studied [24, 25] contained about half of what is presented here. In addition, our analysis is complemented by an extensive literature search of past experimental work on the Alu-associated binding sites that we found, as well as the biological functions of the corresponding TFs and target genes. We started by searching the Alu consensus sequences and found new, previously unnoticed BSs, many of which turned out to be related to developmental processes.
The regulatory role played by Alu for genes that are related to differentiation or development was already established for at least six genes . In these cases Alu was shown to act as enhancer or silencer involved in regulating the expression of six developmental genes. For one of these, the T-cell marker gene CD8α, the silencing mechanism was shown to act via binding of LYF-1, bHLH and GATA3 to the Alu situated in the last intron of CD8α [27, 28]. These experimental results demonstrated the regulatory impact of Alu on developmental genes and during developmental stages. To explore further and enhance the possible regulatory roles of Alu we focused on the genomic region most likely to have a regulatory function - the upstream regions of genes. We scanned these regions and searched for BSs on the particular Alu sequences that were identified. Our findings show for the first time that TFs that are known to regulate the developmental processes may bind to Alu elements that are incorporated in the promoter regions of genes that need to be activated or suppressed during development (such binding has already been demonstrated for Gfi-1, PITX2 and Nkx2.5 - see below).