Poster #RP104
Visual Genomics Methods: Gigantic palindromes and proteins clusters predictions
Sergei Larionov*, Alexander Loskutov*, Eugeny Ryadchenko*, Sergei Rybalko*
*Moscow State University, Moscow, Russia
Based on representation of the DNA sequence as a 2D walk, we consider an
approach to the problem of identification and comparisons of functional and
structural organization of whole chromosomes of different organisms.
According to characteristic design of 2D walks we may identify telomere
sites, palyndromes of various size and complexity, areas of ribosomal RNA ,
transposons, as well as diverse satellite sequences.
Special interest of our research is a gigantic palindromes that we found
in a big amount in different genomes. These palindromes sizes are
in wide intervals : from 5-10 kbs to several hundreds kilobases and may
be more. Most of these sequences have a large part of mutations and looks
like pseudorandom . Alighment of these regions between sequences that compose palindromes give us a low similarity, so it can be defined only by 2D DNA walk method and seems like "context homology" sequences. We tested these sites on KEGGs database and found that some of proteins included in such palindromes organized in functional clusters. In our opinion, such qualitative analysis will provide
an opportunity for a more careful development of identification algorithms,
thereby increasing their predictive capacity, and for discovery of yet
unknown sequence properties.