Jesus E. Garcıa 1, V.A. Gonzalez-Lopez 2, D.S. Santos 3
Department of Statistics, Institute of Mathematics, Statistics and Scientific Computing, University of Campinas, Sergio Buarque de Holanda, 651, Campinas, S.P., CEP: 13083-859, Brazil
Abstract: In this paper, we investigate recently developed proximity measures within the framework of discrete-time Markov processes on discrete state spaces to assess, from multiple perspectives, the similarity and divergence among four Dengue Virus Type 1 genomic sequences available in FASTA format: LC011945, LC011948, LC011949, and LC016760, [1]. It has been speculated that these sequences may originate from multiple sources of serotype 1 of the virus, potentially leading to measurable discrepancies among them, [2]. Each sequence is modeled as a realization of a discrete-time Markov process of order 3 with alphabet {a, c, g, t}. Our analysis identifies LC011949 as the most dissimilar sequence.
Building on this finding, we further identify the specific states most responsible for the observed discrepancies: taa (associated to a stop codon governing translation termination) and atc (related to isoleucine), in that order. We demonstrate that the proximity measures employed are effective in detecting subtle differences that are not immediately evident from the transition probabilities, which are also presented in our study. Based on this analysis, we propose a clustering structure: one group comprising sequences LC011945, LC011948, and LC016760, and a second group consisting solely of LC011949.
© 2026 European Society of Computational Methods in Sciences and Engineering
Keywords: Statistical consistency; Measures between samples; Bayesian information criterion, Efficient determination criterion
Mathematics Subject Classification: 60J10; 60J20