Abstract—Many currently used algorithms for protein coding sequences require large learning sets consisting of real genes to estimate sensible values for used parameters and make the prediction reasonable. They also fail in recognition of short genes because their sequencescontain usually very weak coding signal. To overcome these problems, we worked out a new algorithm for finding protein coding potential in prokaryotic genomes. This algorithm uses homogeneous Markov chain for modeling nucleotide transition between fixed positions in codons thereby reduces the order of Markov chain retaining simultaneously information on dependence between nucleotides in sequence on relatively long distances. We tested performance of this algorithm in relationship to size of the learning set calculating true and false positive rates for different model orders. We also made some comparisons between our algorithm and commonly used GeneMark. The presented algorithm seems to work better than GeneMark especially for smaller learning sets.

Index Terms—Gene finding, markovchains, ORF, protein coding sequence

The authors are with the Department of Genomics, Faculty of Biotechnology, University of Wrocław, ul. Przybyszewskiego 63/77, 51-148 Wrocław, Poland (tel.: +48-071-3756-303; fax: +48-71-3756-234; e-mail: blazej@smorfland.uni.wroc.pl, pamac@smorfland.uni.wroc.pl, cebrat@ smorfland.uni.wroc.pl).

Cite: Paweł Błażej, Paweł Mackiewicz, and Stanisław Cebrat, "New Approach in Coding Signal Recognition Using Homogeneous Markov Chains Independently for Three Codon Positions," International Journal of Bioscience, Biochemistry and Bioinformatics vol. 2, no. 3, pp.  183-187, 2012.