Algorithms for CpG Islands Search: New Advantages and Old Problems

摘要: CpG islands (CGIs) are regions having high GC and content while generally mammalian genomes CpG-depleted. CGIs often located in the promoter region of genes, mostly housekeeping but also tissue-specific. It is widely believed that dinucleotides within promoters unmethylated targets for specific regulatory protein binding. As a result, contain special sequence motifs highly affinitive binding (transcription factor sites, TFBS). Methylation cytosine context such could decrease affinity TF binding, increase attraction methyl-binding proteins, affect histones modification and, therefore, leads to repression genes transcription. The mechanism local global transcription via methylation used many different normal (development, differentiation, aging, X-chromosome inactivation, imprinting) pathological processes (cancer other diseases). However recently it has been reported class normally methylated active do exist. Lately evidences biological relevance or far from gene appear. Such act as regulator pervasive transcription, which seems be actual genome feature rather than side-effect high-throughput techniques errors. Replication origins associated with any location. consequence nucleotide content, DNA RNA secondary structures. For example, G2-3C2-3 motif common induces significant curiosity DNA. Another motif, G-rich (GRS) 3’ 5’ RNA, known form structures, G-quadruplexes, on both end playing important role its stability. This corresponds C-rich DNA, likely appear CGIs. Classical algorithms search use sliding window (SWM) running sum (RSM) several distinct not independent criteria (GC Obs/ExpCpG length). thresholds arbitrary, unconcerned between species, demonstrate lack interpretation. SWM slow, RSM faster tend split large into smaller ones omit nonuniform distribution along sequence. Recently, based clustering were implemented. Those have number parameters reasonable mathematical basics. comparison tricky. Hypermutability lead loss