γ-MYN: a new algorithm for estimating Ka and Ks with consideration of variable substitution rates

• Published in: Biology direct Vol. 4; no. 1; p. 20
• Main Authors: Wang, Da-Peng, Wan, Hao-Lei, Zhang, Song, Yu, Jun
• Format: Journal Article
• Language: English
• Published: London BioMed Central 16.06.2009; BioMed Central Ltd; BMC
• Subjects: Algorithms; Amino Acid Substitution - genetics; Animals; Approximation theory; Base Sequence; Biomedical and Life Sciences; Codon - genetics; Computational Biology - methods; Computer Simulation; DNA sequencing; Gene mutations; Genetic algorithms; Humans; Identification and classification; Life Sciences; Nucleotide sequence; Nucleotide sequencing; Selection, Genetic; Time Factors; Substitution Rate; Transition Probability Matrix; Negative Selection; Gamma Distribution; Nucleotide Frequency
• ISSN: 1745-6150; 1745-6150
• DOI: 10.1186/1745-6150-4-20

Background Over the past two decades, there have been several approximate methods that adopt different mutation models and used for estimating nonsynonymous and synonymous substitution rates (Ka and Ks) based on protein-coding sequences across species or even different evolutionary lineages. Among them, MYN method (a M odified version of Y ang- N ielsen method) considers three major dynamic features of evolving DNA sequences–bias in transition/transversion rate, nucleotide frequency, and unequal transitional substitution but leaves out another important feature: unequal substitution rates among different sites or nucleotide positions. Results We incorporated a new feature for analyzing evolving DNA sequences–unequal substitution rates among different sites–into MYN method, and proposed a modified version, namely γ (gamma)-MYN, based on an assumption that the evolutionary rate at each site follows a mode of γ -distribution. We applied γ -MYN to analyze the key estimator of selective pressure ω (Ka/Ks) and other relevant parameters in comparison to two other related methods, YN and MYN, and found that neglecting the variation of substitution rates among different sites may lead to biased estimations of ω. Our new method appears to have minimal deviations when relevant parameters vary within normal ranges defined by empirical data. Conclusion Our results indicate that unequal substitution rates among different sites have variable influences on ω under different evolutionary rates while both transition/transversion rate ratio and unequal nucleotide frequencies affect Ka and Ks thus selective pressure ω. Reviewers This paper was reviewed by Kateryna Makova, David A. Liberles (nominated by David H Ardell), Zhaolei Zhang (nominated by Mark Gerstein), and Shamil Sunyaev.