Beyond the Identification of Transcribed Sequences:
Functional, Evolutionary and Expression Analysis
12th International Workshop
October 25-28, 2002
Washington, DC

List of Abstracts * Speakers * Organizers * Authors * Original Announcement

Statistical Validation of Pattern Prevalence in Regulatory Gene Loci

Gautam B. Singh*  and Stephen A. Krawetz**       
*Computer Science & Engineering Department, Oakland University, Rochester, MI 48309, USA; FutureSoft Corporation, Rochester Hills, MI 48309, USA; Michigan Center for Biological Information, WSU Node
** Michigan Center for Biological Information, WSU Node, Wayne State University School of Medicine, Detroit, MI 48202

Gene expression in cells is controlled by the presence of DNA sequences called Matrix Association Regions (MARs) or Scaffold Association Regions (SARs). These sequences are embedded within the majority of nuclear non-coding DNA.  MARs regions are found to be distributed throughout the chromatin and flank the ends of genic domains encompassing various transcriptional units. It has also been shown that MARs bring together the transcriptionally active regions of chromatin such that transcription is initiated in the region of the chromosome that coincides with the surface of the nuclear matrix. Specific patterns of nucleotide arrangements have been identified in these non-coding DNA sequences, with varying functions. These examples indicate that the patterns embedded in the eukaryote DNA play an important role in its viability. Examples of these patterns include the A+T or G+C rich regions, telomeric repeats of AGGGTTin human DNA, the rare occurrence or absence of dinucleotides, e.g., TA and GC, and tetranucleotides, e.g., CTAG, and the GNN periodicity in the gene coding regions.

In this talk we will discuss the prevalence of MAR specific patterns with in the set of sequences collected within the SMAR-DB, as well as an independently collected set of sequences from various Origins of Replication (ORI). Interesting statistics related to curved, kinked, AT-rich regions, and other MAR-related patterns are presented. The actual occurrence of patterns is compared with their expected occurrence with the result of the log-ratio test used for establishing whether the putative patterns are an evidence of the occurrence of MARs. The patterns that are positively linked with MARs are further analyzed for periodicity. This is the first step towards testing the hypothesis that MAR periodicity may be associated with DNA packaging. The statistical analysis is presented within the Bayesian context.

A MAR-Pattern Distributed Annotation Sources (DAS) will be linked to the ENSEMBL database as part of the Michigan Center for Biological Information (MCBI) Genome databases.  This will provide MAR-related pattern loci information within the context of the existing annotations that are targeted towards gene transcription. This will hopefully help us better appreciate the correlation between the transcription apparatus and the DNA structure. 

  Abstract List

List of Abstracts * Speakers * Organizers * Authors * Original Announcement

Genetic Meetings