Key-string algorithm--novel approach to computational analysis of repetitive sequences in human centromeric DNA.

AIM To use a novel computational approach, Key-string Algorithm (KSA), for the identification and analysis of arbitrarily large repetitive sequences and higher-order repeats (HORs) in noncoding DNA. This approach is based on the use of key string that plays a role of an arbitrarily constructed "computer enzyme". METHOD A cluster of novel KSA-related methods was introduced and developed on the basis of a combination of computations on a very modest scale, by eye inspection and graphical display of results of analysis. Sequence analysis software was developed, containing seven programs for KSA-related analyses. This approach was demonstrated in the case study of alpha satellites and HORs in the human genetic sequence AC017075.8 (193277 bp) from the centromeric region of human chromosome 7. The KSA segmentation method was applied by using DCCGTTT, GTA, and TTTC key strings. RESULTS Fifty-five copies of 2734-bp 16mer HORs were identified and investigated, and a start-string TTTTTTAAAAA was identified. The HOR-matrix was constructed and employed for graphical display of mutations. KSA identification of HORs in AC017075.8 was compared with that of RepeatMasker and Tandem Repeat Finder, which identified alpha monomers in AC017075.8, but not the HORs. On the basis of KSA study, the centromere folding was described as an effect of HORs and super-HORs (3 x 2734 bp) in AC017075.8. The following novel computational KSA-based methods, easy-to-use and intended for computational "pedestrians", were demonstrated: color-HOR diagram, KSA-divergence method, 171-bp subsequence-convergence diagram, and total frequency distribution of the key-string subsequence lengths. The results were supplemented by Fast Fourier Transform, employing a novel mapping of symbolic genomic sequence into a numerical sequence. CONCLUSION The KSA approach offers a simple and robust framework for a wide range of investigations of large repetitive sequences and HORs, involving a very modest scope of computations that can be carried out by using a PC. As the KSA method is HOR-oriented, the identification of HORs is even easier than the identification of underlying alpha monomer itself. This approach provides an easy identification of point mutations, insertions, and deletions, with respect to consensus. This may be useful in a wide range of investigations and applied in forensic medicine, medical diagnosis of malignant diseases, biological evolution, and paleontology.