Bioinformatics Advance Access originally published online on October 12, 2004
Bioinformatics 2005 21(6):811-816; doi:10.1093/bioinformatics/bti059
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TRbase: a database relating tandem repeats to disease genes for the human genome
School of Biological and Chemical Sciences, Washington Singer Laboratories, University of Exeter Perry Road, Exeter EX4 4QG, UK
*To whom correspondence should be addressed.
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONIMPLEMENTATION AND RESULTSCONCLUSIONS AND FUTURE...REFERENCES |
|---|
Motivation: Tandem repeats are associated with disease genes, play an important role in evolution and are important in genomic organization and function. Although much research has been done on short perfect patterns of repeats, there has been less focus on imperfect repeats. Thus, there is an acute need for a tandem repeats database that provides reliable and up to date information on both perfect and imperfect tandem repeats in the human genome and relates these to disease genes.
Results: This paper presents a web-accessible relational tandem repeats database that relates tandem repeats to gene locations and disease genes of the human genome. In contrast to other available databases, this database identifies both perfect and imperfect repeats of 1–2000 bp unit lengths. The utility of this database has been illustrated by analysing these repeats for their distribution and frequencies across chromosomes and genomic locations and between protein-coding and non-coding regions. The applicability of this database to identify diseases associated with previously uncharacterized tandem repeats is demonstrated.
Availability: TRbase is available at http://trbase.ex.ac.uk/
Contact: S.J.Aves{at}exeter.ac.uk
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONIMPLEMENTATION AND RESULTSCONCLUSIONS AND FUTURE...REFERENCES |
|---|
Tandem repeats have been identified and are widespread in eukaryotic genomes from yeast to humans (Vergnaud and Denoeud, 2000; Tóth et al., 2000; Katti et al., 2001). A tandem repeat consists of head-to-tail copies of a nucleotide sequence that may be exact or, more commonly, approximate. The propensity for tandemly repeated sequences to expand and contract results in copy number polymorphisms, and they have been extensively exploited as genetic markers for mapping, population and DNA fingerprinting studies (Calafell et al., 1998). Despite this wide exploitation, the mechanisms that generate and maintain tandem repeats in genomes are poorly understood. Moreover, there is accumulating evidence that tandem repeats may serve functional roles within coding sequences and as regulatory elements or mutational hotspots (Kashi et al., 1997).
Expansions in tandem repeats have been associated with human disease, including neurodegenerative disorders such as fragile X syndrome, Huntington’s disease and spinocerebellar ataxia, and some cancers for example hereditary nonpolyposis colorectal carcinoma (Sutherland and Richards, 1995; Mitas, 1997; Tóth et al., 2000). These tandem repeat expansions may occur in protein-coding regions of genes (Machado–Joseph disease, Haw River syndrome, Huntington's disease) or non-coding regions (Friedrich ataxia, myotonic dystrophy, some forms of fragile X syndrome). Expansions involving minisatellite, pentanucleotide, tetranucleotide and numerous trinucleotide repeats have been associated with diseases or fragile sites (Richard and Pâques, 2000; Ranum and Day, 2002; http://www.neuro.wustl.edu/neuromuscular/mother/dnarep.htm). The expansions associated with these diseases have been proposed to be caused by replication slippage or asymmetric recombination; however, the exact mechanisms responsible for tandem repeat expansion are still controversial (Chambers and MacAvoy, 2000; Richard and Pâques, 2000; Viguera et al., 2001).
Perfect tandem repeats are relatively simple to detect, and many studies have characterized such perfect repeats in a variety of sequenced genomes (e.g. Yeramian and Buc, 1999; Tóth et al. 2000; le Flèche et al. 2001; Katti et al. 2001; Subramanian et al. 2003). Such studies are limited, however, because genomic tandem repeats are seldom perfect copies: mutation events can introduce mismatches or indels (insertions or deletions) into copies of the repeated sequence, with subsequent expansion and contraction giving rise to complex tandem repeat patterns (Benson, 2002). Such ‘real’ tandem repeats are difficult to detect using standard algorithms. Recently, specialist software programs for the detection of genomic tandem repeats has been developed, including the Tandem Repeats Finder (TRF) program, which is an efficient and sensitive tool for detecting tandem repeats that have undergone extensive mutational change (Benson, 1999) STRING (Parisi et al., 2003) STAR (Delgrange and Rivals, 2004) and an exhaustive whole-genome search algorithm (Krishnan and Tang, 2004).
Many databases exist of perfect tandem repeats, but the focus on short perfect repeats has left gaps in our understanding of the potentially important biological or medical roles of those that are longer and harder to detect. In this study, we describe a database of perfect and imperfect (‘real’) tandem repeats in chromosomes of the human genome and demonstrate how this allows straightforward but flexible searches for tandem repeats in combination with other important features including disease loci.
|
IMPLEMENTATION AND RESULTS |
|---|
|
TOPAbstractINTRODUCTIONIMPLEMENTATION AND RESULTSCONCLUSIONS AND FUTURE...REFERENCES |
|---|
TRbase structure and content
TRbase is a publicly available relational database of tandem repeats in the human genome that has been developed using MySQL version 4.0 (www.mysql.com). We detected tandem repeats using the TRF program (version 3.01; Benson, 1999) applied to DNA sequences extracted from GenBank in the FASTA format using the Seqret program of EMBOSS (Rice et al., 2000). DNA sequences and annotations were retrieved for the completed chromosomes 4, 5, 6, 14, 16, 18, 19, 20, 21 and 22 (October 2003, Build number 34). TRF has several major advantages compared with other tandem repeat detection algorithms: it is very flexible, being able to detect repeats of any size; no prior knowledge of repeat patterns or pattern size is required; and, importantly, it is not restricted to perfect repeats but can detect tandem repeats that have accumulated extensive mutational change, including insertions or deletions. To allow users of TRbase flexibility in their searches, we employed both high- and low-stringency TRF detection parameters (match, mismatch and insertion/deletion scores of +2,–7,–7 or +2,–5,–5, respectively) with a minimal alignment score of 45 as the cut-off for reporting repeats (Benson, 1999). The high-stringency dataset is appropriate for many tandem repeat searches, including perfect or near-perfect repeats, and therefore this was set as the default in TRbase; users may choose the low-stringency option if they wish to detect tandem repeats that have undergone extensive mutational change. Further search refinement is possible within TRbase by defining the percentage mismatch allowed between repeats (see below).
The central table of TRbase contains all the tandem repeat data precomputed using the TRF program. This is linked to a gene features table via a linking table in which start and end positions of the tandem repeats are mapped to the start and end positions of exons, introns and intergene regions. Tandem repeats are classified as fully ‘within’ an exon, intron or intergene, ‘overlapping’ a boundary, or as ‘spanning’ two or more boundaries. An important feature of TRbase is to relate disease information to tandem repeats. Therefore, the tandem repeat table is also linked to a table containing information on disease genes on all 24 chromosomes of the human genome. Disease information was downloaded from the Online Mendelian Inheritance in Man (OMIM) database, which provides bibliographic, sequence and map information on genetic disorders, including trinucleotide expansion disorders (Hamosh et al., 2002). To automate the process of creating the tables and populating the database, all SQL queries were embedded and implemented using the universal PERL DBI module.
TRbase can be searched for information by means of Web pages dynamically generated using PHP 3.0 scripts. The main search page allows a search across all the three main information areas and returns information on the tandem repeats identified within the prescribed parameters and any related gene names or disease details, with their links to the NCBI or OMIM databases, respectively. Further search pages allow more detailed information to be retrieved or more complex searches to be performed; searches for a repeat unit consensus sequence will return hits for all possible permutations of the repeated consensus and its complement. Links are provided to other tandem repeat databases including the ABCC GRID database (Collins et al., 2003), Minisatellite database (le Flèche et al., 2001), PlantSat database (Macas et al., 2002), RSDB (Horng et al., 2002), the Microsatellite Analysis Server (MICAS) for prokaryotic genomes (Sreenu et al., 2003) and the Tandem Repeats Finder homepage (Benson, 1999).
Comparison of TRbase with other databases
Most existing tandem repeats databases are limited as they focus on microsatellites or perfect repeats, because these are easiest to detect. For example, the GRID database of microsatellites (Collins et al., 2003) has considered only perfect repeats of 2–16 bp units. As the number of imperfect repeats stored in our database is almost four times the number of perfect repeats, the GRID database is severely limited in the number of repeats identified (Table 1). We also compared TRbase to the Minisatellite database (le Flèche et al., 2001). Although currently limited to chromosomes 20, 21 and 22 of the human genome, the Minisatellite database is similar to TRbase in that it stores tandem repeats computed by TRF. However, discrepancies were found in the tandem repeats identified (Table 1). Although the total number of repeats found in the Minisatellite database is more than that present in the TRbase, many perfect repeats present in the TRbase are not picked up by the Minisatellite database (Table 1), but are replaced by a longer imperfect repeat. This may be due to the fact that the Minisatellite database has set the TRF parameters to a much lower stringency, and has therefore identified longer repeats with lower scores that overlap perfect repeats. Such repeats may thus need to be used with caution. Importantly, TRbase permits searches not available with the Minisatellite database including on disease genes and gene features (see below), as well as giving links to external databases like Ensembl, NCBI and OMIM.
|
Tandem repeat statistics
Using high-stringency search parameters (match, mismatch and insertion/deletion scores of +2,–7,–7, respectively; minimal alignment score
45; no additional restrictions on repeat unit length or total length), the density of all tandem repeats averages 295 per megabase of genomic DNA, or one tandem repeat per 3.4 kb. These form 21 670 bp per megabase or 2.2% of the genome. The maximum density (446 Mb–1; 5.0%) is found in chromosome 19, with the minimum in chromosome 4 (264 Mb– 1; 1.6%). Chromosome 19 has a high frequency of genes (Lander et al., 2001) and we find that these have a much higher frequency of tandem repeats than genes in other chromosomes: 390 Mb– 1 compared with an average of 164 Mb–1. For all chromosomes analysed, tandem repeat frequencies decrease markedly with increasing repeat unit length, copy number and repeat lengths (Fig. 1), although it is notable that there are nearly as many repeats with copy numbers 21–30 as those with copy numbers between 11 and 20 (Fig. 1C). Most chromosomes show similar trends but chromosome 19 has the maximum number of repeats at higher repeat unit lengths (Fig. 1A), in line with its higher content of exonic tandem repeats.
|
Tandem repeats were predominantly found in intergenic and intronic regions: only 0.66% of all repeats were present within exons. As exons represent 1.18% of genomic sequence, tandem repeats are therefore underrepresented in exons, although tandem repeats in exons tend to be longer (mean total length = 216 versus 84 bp for tandem repeats in introns and intergenes). It was found that 21.7% of genes contained at least one tandem repeat. To determine the frequencies of different repeat types with respect to their location in genes, repeats with unit length of 1–30 bp were analysed in more detail (Fig. 2). These repeats were not uniformly distributed: repeats with a motif length divisible by three were greatly enriched in exons compared with introns and intergenic regions (Fig. 2). The presence of such repeats in coding regions would not lead to frameshift mutations, thus preserving protein structure and functionality (Metzgar et al., 2001; Bor
tnik and Pumpernik, 2002. Figure 2 shows that the distributions of tandem repeat types in non-coding regions (introns and intergenes) are very similar to each other, with a great predominance of microsatellites, and minisatellites of short repeat unit length. The most frequent microsatellite repeats are dinucleotide repeats: the most common consensus sequence is AC and there are no CG repeats, confirming the findings of Lander et al., (2001). Suppression of CG dinucleotides has been observed in many eukaryotic genomes (Katti et al., 2001) and may be due to methylation of cytosine which increases its chances of mutation to thymine by deamination (Schorderet and Gartler, 1992). Our observed abundance of AC repeats is interesting as Majewski and Ott, (2000) have shown that there is a significant correlation between such repeats and recombination hot spots. There is a relative paucity of trinucleotide repeats in non-coding regions, contrasting with their abundance in coding regions. Our tandem repeat data in Figure 2 display very similar patterns to an analysis restricted to perfect microsatellites of short repeat unit length (Subramanian et al., 2003).
|
30 in exon, intron and intergenic regions. Percentages are of the total number of tandem repeats in each class: exons, black bars; introns, grey bars; and intergenes, white bars. Detection parameters were as for Using TRbase to detect tandem repeats associated with disease genes
A key aim of TRbase was to facilitate searches for tandem repeats associated with disease genes. Table 2 shows data from a high-stringency TRbase search for trinucleotide repeats in exons of disease genes. Trinucleotide repeat expansions have been shown to cause several human diseases, therefore it is of considerable interest to identify such repeats in disease genes. From Table 2 it can be seen that the known exonic disease-causing triplet repeats have been identified in the genes CACNA1A (migraine), SCA1 (spinocerebellar ataxia 1), TBP (spinocerebellar ataxia 17) and HD (Huntington's disease). In addition, this search reveals a further 90 trinucleotide repeats in exons of disease genes, including 33 repeats with percentage match scores of at least 90% (Table 2). Therefore, TRbase is a useful tool for identifying both perfect and imperfect tandem repeats associated with disease genes, further study of which may be informative about the genetic basis of disease. For example, Table 2 shows that the RUNX2 gene contains 22 copies of a CAG repeat. Similar to the known disease-causing triplet repeats shown in boldface in Table 2 this encodes a polyglutamine tract and is subject to length polymorphisms (Vaughan et al., 2002). However, while mutations in the RUNX2 (CBFA1) gene can give rise to cleidocranial dysplasia, CAG repeat variants may result in other phenotypes: Vaughan et al., (2002) concluded from a study of 495 Australian women that they may be correlated with reduced bone mineral density and bone fracture. Table 1 also shows that RUNX2 gene contains a second, shorter trinucleotide repeat, which encodes polyalanine. This also exhibits polymorphism but does not appear to be related to bone mineral density, although further analysis in other populations is required (Vaughan et al., 2002).
|
As a further example, Table 2 shows that the HD gene contains two trinucleotide repeat sequences: the CAG repeat known to be associated with Huntington's disease and an adjacent CCG triplet repeat coding for a stretch of proline residues (shown as GCC consensus in Table 2). This CCG triplet repeat has implications for diagnostic testing as it is also known to be polymorphic (Andrew et al., 1994), and a recent multiple regression analysis of polymorphisms in HD gene has indicated that variation in the CCG repeats might modify the age of onset in Huntington’s disease (Chattopadhyay et al., 2003) although this remains to be demonstrated in all populations (Wang et al., 2004).
|
CONCLUSIONS AND FUTURE PERSPECTIVE |
|---|
|
TOPAbstractINTRODUCTIONIMPLEMENTATION AND RESULTSCONCLUSIONS AND FUTURE...REFERENCES |
|---|
TRbase provides a platform to study the associations between disease genes and previously uncharacterized tandem repeats which is a key to understanding the mechanisms of pathogenicity of such diseases. These associations, together with the data indicating distribution patterns in protein-coding and non-coding regions could help direct future research into identifying the evolutionary history and functional significance of tandem repeats. Currently, the database contains tandem repeat information linked to gene positions and disease genes for the 10 most complete human chromosomes; database creation has been automated and the database will be updated as further refined builds for the remaining chromosomes are released. Further investigations are being continued for incorporating information on tandem repeat polymorphisms and links to the SNP database.
|
Acknowledgments |
|---|
We would like to thank Robin Batten, Andrew Dalby, Matthew Redden and Darren Soanes for their valuable suggestions and assistance. A.M.P. is funded by a studentship from the UK Biotechnology and Biological Sciences Research Council.
Received on April 1, 2004; revised on August 21, 2004; accepted on September 20, 2004
|
REFERENCES |
|---|
|
TOPAbstractINTRODUCTIONIMPLEMENTATION AND RESULTSCONCLUSIONS AND FUTURE...REFERENCES |
|---|
Andrew, S.E., Goldberg, Y.P., Theilmann, J., Zeisler, J., Hayden, M.R. (1994) A CCG repeat polymorphism adjacent to the CAG repeat in the Huntington's disease gene: implications for diagnostic accuracy and predictive testing. Hum. Mol. Genet., 3, 65–67
Benson, G. (1999) Tandem Repeats Finder: a program to analyze DNA sequences. Nucleic Acids Res., 27, 573–580
Benson, G. (2002) A new distance measure for comparing sequence profiles based on path lengths along an entropy surface. Bioinformatics, 18, (Suppl. 2), S44–S53[Abstract].
Bortnik, B. and Pumpernik, D. (2002) Tandem repeats in protein coding regions of primate genes. Genome Res., 12, 909–915
Calafell, F., Shuster, A., Speed, W.C., Kidd, J.R., Kidd, K.K. (1998) Short tandem repeat polymorphism evolution in humans. Eur. J. Hum. Genet., 6, 38–49[CrossRef][Web of Science][Medline].
Chambers, G.K. and MacAvoy, E.S. (2000) Microsatellites: consensus and controversy. Comp. Biochem. Physiol. B Biochem. Mol. Biol., 126, 455–476[CrossRef][Medline].
Chattopadhyay, B., Ghosh, S., Gangopadhyay, P.K., Das, S.K., Roy, T., Sinha, K.K., Jha, D.K., Mukherjee, S.C., Chakraborty, A., Singhal, B.S., Bhattacharya, A.K., Bhattacharyya, N.P. (2003) Modulation of age at onset in Huntington's disease and spinocerebellar ataxia type 2 patients originated from eastern India. Neurosci. Lett., 345, 93–96[CrossRef][Web of Science][Medline].
Collins, J.R., Stephens, R.M., Gold, B., Long, B., Dean, M., Burt, S.K. (2003) An exhaustive DNA micro-satellite map of the human genome using high performance computing. Genomics, 82, 10–19[CrossRef][Web of Science][Medline].
Delgrange, O. and Rivals, E. (2004) STAR: an algorithm to search for tandem approximate repeats. Bioinformatics, 20, 2812–2820
Hamosh, A., Scott, A.F., Amberger, J., Bocchini, C., Valle, D., McKusick, V.A. (2002) Online Mendelian Inheritance in Man (OMIM), a knowledgebase of human genes and genetic disorders. Nucleic Acids Res., 30, 52–55
Horng, J.T., Huang, H.D., Jin, M.H., Wu, L.C., Huang, S.L. (2002) The repetitive sequence database and mining putative regulatory elements in gene promoter regions. J. Comput. Biol., 9, 621–640[CrossRef][Web of Science][Medline].
Kashi, Y., King, D., Soller, M. (1997) Simple sequence repeats as a source of quantitative genetic variation. Trends Genet, 13, 74–78[CrossRef][Web of Science][Medline].
Katti, M.V., Ranjekar, P.K., Gupta, V.S. (2001) Differential distribution of simple sequence repeats in eukaryotic genome sequences. Mol. Biol. Evol., 18, 1161–1167
Krishnan, A. and Tang, F. (2004) Exhaustive whole-genome tandem repeats search. Bioinformatics, 20, 2702–2710
Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al. (2001) Initial sequencing and analysis of the human genome. Nature, 409, 860–921[CrossRef][Medline].
le Flèche, P., Hauck, Y., Onteniente, L., Prieur, A., Denoeud, F., Ramisse, V., Sylvestre, P., Benson, G., Ramisse, F., Vergnaud, G. (2001) A tandem repeats database for bacterial genomes: application to the genotyping of Yersinia pestis and Bacillus anthracis. BMC Microbiol, 1, 2[CrossRef][Medline].
Macas, J., Meszaros, T., Nouzova, M. (2002) PlantSat: a specialized database for plant satellite repeats. Bioinformatics, 18, 28–35
Majewski, J. and Ott, J. (2000) GT repeats are associated with recombination on human chromosome 22. Genome Res., 10, 1108–1114
Metzgar, D., Thomas, E., Davis, C., Field, D., Wills, C. (2001) The microsatellites of Escherichia coli: rapidly evolving repetitive DNAs in a non-pathogenic prokaryote. Mol. Microbiol., 39, 183–190[CrossRef][Web of Science][Medline].
Mitas, M. (1997) Trinucleotide repeats associated with human disease. Nucleic Acids Res., 25, 2245–2254
Parisi, V., De Fonzo, V., Aluffi-Pentini, A. (2003) STRING: finding tandem repeats in DNA sequences. Bioinformatics, 19, 1733–1738
Ranum, L.P. and Day, J.W. (2002) Dominantly inherited, non-coding microsatellite expansion disorders. Curr. Opin. Genet. Dev., 12, 266–271[CrossRef][Web of Science][Medline].
Rice, P., Longden, I., Bleasby, A. (2000) EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet., 16, 276–277[CrossRef][Web of Science][Medline].
Richard, G.F. and Pâques, F. (2000) Mini- and microsatellite expansions: the recombination connection. EMBO Rep., 1, 122–126[CrossRef][Web of Science][Medline].
Schorderet, D.F. and Gartler, S.M. (1992) Analysis of CpG suppression in methylated and nonmethylated species. Proc. Natl Acad. Sci. USA, 89, 957–961
Sreenu, V.B., Alevoor, V., Nagaraju, J., Nagarajaram, H.A. (2003) MICdb: database of prokaryotic microsatellites. Nucleic Acids Res., 31, 106–108
Subramanian, S., Mishra, R.K., Singh, L. (2003) Genome-wide analysis of microsatellite repeats in humans: their abundance and density in specific genomic regions. Genome Biol., 4, R13[CrossRef][Medline].
Sutherland, G.R. and Richards, R.I. (1995) Simple tandem DNA repeats and human genetic disease. Proc. Natl Acad. Sci. USA, 92, 3636–3641
Tóth, G., Gaspari, Z., Jurka, J. (2000) Microsatellites in different eukaryotic genomes: survey and analysis. Genome Res., 10, 967–981
Vaughan, T., Pasco, J.A., Kotowicz, M.A., Nicholson, G.C., Morrison, N.A. (2002) Alleles of RUNX2/CBFA1 gene are associated with differences in bone mineral density and risk of fracture. J. Bone Miner. Res., 17, 1527–1534[CrossRef][Web of Science][Medline].
Vergnaud, G. and Denoeud, F. (2000) Minisatellites: mutability and genome architecture. Genome Res., 10, 899–907
Viguera, E., Canceill, D., Ehrlich, S.D. (2001) Replication slippage involves DNA polymerase pausing and dissociation. EMBO J., 20, 2587–2595[CrossRef][Web of Science][Medline].
Wang, C.K., Wu, Y.R., Hwu, W.L., Chen, C.M., Ro, L.S., Chen, S.T., Gwinn-Hardy, K., Yang, C.C., Wu, R.M., Chen, T.F., Wang, H.C., Chao, M.C., Chiu, M.J., Lu, C.J., Lee-Chen, G.J. (2004) DNA haplotype analysis of CAG repeat in Taiwanese Huntington's disease patients. Eur. Neurol., 52, 96–100[CrossRef][Web of Science][Medline].
Yeramian, E. and Buc, H. (1999) Tandem repeats in complete bacterial genome sequences: sequence and structural analyses for comparative studies. Res. Microbiol., 150, 745–754[Medline].
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Merkel and N. Gemmell Detecting short tandem repeats from genome data: opening the software black box Brief Bioinform, September 1, 2008; 9(5): 355 - 366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Ames, N. Murphy, T. Helentjaris, N. Sun, and V. Chandler Comparative Analyses of Human Single- and Multilocus Tandem Repeats Genetics, July 1, 2008; 179(3): 1693 - 1704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. B. Mudunuri and H. A. Nagarajaram IMEx: Imperfect Microsatellite Extractor Bioinformatics, May 15, 2007; 23(10): 1181 - 1187. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.-M. Patch and S. J. Aves Fingerprinting fission yeast: polymorphic markers for molecular genetic analysis of Schizosaccharomyces pombe strains Microbiology, March 1, 2007; 153(3): 887 - 897. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Shelenkov, K. Skryabin, and E. Korotkov Search and Classification of Potential Minisatellite Sequences from Bacterial Genomes DNA Res, January 1, 2006; 13(3): 89 - 102. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||