Abstract: 【Objective】 The quality of Paeonia lactiflora flower color affects its ornamental value and the commercial value of ornamental plants. This study aims to understand the codon usage pattern of genes regulating flower color and probe into the main factors affecting the formation of codon bias, which has important biological significance for mRNA translation, design of transgenes, the prediction of expression level and functions of new genes, and studies of molecular biology and evolution, etc. 【Method】In a previous study, 6,345 differential genes were screened out by transcriptome sequencing of a flower color chimaera cultivar “Jinhui” with a consistent genetic background red outer-petal and yellow inner-petal, followed by a further filtering analysis according to the principle of CDS sequence characteristics and greater than 300 bp. We finally obtained 2,234 genes as our research object. Mobyle software was used to calculate different parameters for the codon usage, such as GC content, average GC content of the first and second positions (GC12), GC content of the third position (GC3s), effective number of codon (ENC), codon adaptation index (CAI), and relative synonymous codon usage (RSCU). Further analysis of a neutrality plot (GC12 vs. GC3), an ENC-GC3s plot, and a Parity Rule 2 (PR2) plot were performed. Additionally, we probed into the influence of mutational pressure and translational selection by a multivariate statistical analysis. Finally, we took 5% CAI value as high-expression and low-expression sample groups, then calculated the RSCU value, and analyzed the significant difference to determine the optimal codons by a chi-square test.【Result】The results showed that the GC content at the third position of codons was 46.37%. The GC content of most genes was mainly distributed between 30% and 55%. Neutrality analysis showed that there was a significant positive correlation (R²=0.202, P＜0.01) between GC3s and GC12 value. The ENC-plot showed most of the genes on or close to the expected curve, but also some points with low-ENC values were below it. The(ENCexp-ENCobs)/ENCexp ratio of most genes ranged from -0.05 to 0.05. The Parity Rule 2-plot showed that the frequency of T nucleotide at the third position was higher than A, and C was higher than G, suggesting that the use frequencies of four nucleotide were not balanced. Correspondence analysis showed that the first axis showed a 38.09% variation, while the other three axes showed 18.42%, 15.09%, and 14.59%, respectively, suggesting that the first axis was the main index evaluating the codon usage bias of Paeonia lactiflora genes regulating flower color. Mutation pressure and selection analysis showed there were significant negative correlations (R²=0.736, P＜0.01. R²=0.286, P＜0.01) between the first axis and GC3s, CAI value, respectively. Using the delta RSCU and significant chi-square test methods, we defined 21 codons as the major preference codons in the Paeonia lactiflora genes regulating flower color, of 18 codons ending with G or C, only CGU and GGU ending with U.【Conclusion】In conclusion, most of optimal codons ended with G or C. Meanwhile, the codon usage pattern of Paeonia lactiflora genes regulating flower color is formed under the effect of mutational pressure (R²=0.736) and translational selection (R²=0.286), but a mutational bias was the major influence on codon usage. This study not only preliminarily reveals the codon usage pattern of Paeonia lactiflora genes, but also provides a certain theoretical basis for further carrying out genetic improvement of Paeonia lactiflora flower color by codon reconstruction and analyzing the molecular evolution.

Key words: Paeony, flower color controlling gene, codon usage pattern, influence factor