Analysis of codon usage patterns in citrus based on coding sequence data

Abstract

Background: Codon usage is an important determinant of gene expression levels that can help us understand codon biology, evolution and mRNA translation of species. The majority of previous codon usage studies have focused on single species analysis, although few studies have focused on the species within the same genus. In this study, we proposed a multispecies codon usage analysis workflow to reveal the genetic features and correlation in citrus.

Results: Our codon usage analysis workflow was based on the GC content, GC plot, and relative synonymous codon usage value of each codon in 8 citrus species. This approach allows for the comparison of codon usage bias of different citrus species. Next, we performed cluster analysis and obtained an overview of the relationship in citrus. However, traditional methods cannot conduct quantitative analysis of the correlation. To further estimate the correlation among the citrus species, we used the frequency profile to construct feature vectors of each species. The Pearson correlation coefficient was used to quantitatively analyze the distance among the citrus species. This result was consistent with the cluster analysis.

Conclusions: Our findings showed that the citrus species are conserved at the genetic level and demonstrated the existing genetic evolutionary relationship in citrus. This work provides new insights into codon biology and the evolution of citrus and other plant species.

Keywords: Citrus; Codon usage; Correlation; Evolution; GC biology.

Fig. 1

Neutrality plot of 8 citrus species. The green solid line represents the regression line. a Atlantia buxifolia, the regression line is y=−0.0258x+46.5950,R²=0.0418. b Fortunella hindsii, the regression line is y=0.0781x+42.5627,R²=0.1218. c Citrus grandis, the regression line is y=0.0921+41.9104,R²=0.1288. d Citrus sinensis, the regression line is y=0.2712x+34.4916,R²=0.3047. e Citrus medica, the regression line is y=−0.0275x+46.6589,R²=0.0494. f Citrus reticulata ‘Mangshan’, the regression line is y=−0.0954x+49.2216,R²=0.1476. g Citrus ichangensis, the regression line is y=0.0174x+44.8579,R²=0.0341. h Citrus clementina, the regression line is y=−0.2456x+51.3338,R²=0.0341

Fig. 2

Neutrality plot of 8 citrus species. ENCs were plotted against GC content at the third position. The green solid line represents the expected curve of positions of genes when the codon usage was only determined by the GC3s composition. a Atlantia buxifolia. b Fortunella hindsii. c Citrus grandis. d Citrus sinensis. e Citrus medica. f Citrus reticulata ‘Mangshan’. g Citrus ichangensis. h Citrus clementina

Fig. 3

Frequency distribution of (ENCexp-ENCobs)/ENCexp. ENCexp represents expected ENC values and ENCobs represents ENC observed values. The peak located in 0 to 0.1

Fig. 4

Heat map of RSCU of 59 codons from 30 species using Euclidean distance and average clustering module. GC and GC3 distribution in ORFs from 30 plant genomes

Fig. 5

Heat map of pearson correlation coefficient among Citrus species. Ab: Atlantia buxifolia; Fh: Fortunella hindsii; Cg: Citrus grandis; Cs: Citrus sinensis; Cm: Citrus medica; Cr: Citrus reticulata ‘Mangshan’; Ci: Citrus ichangensis; Cc: Citrus Clementina

Fig. 6

Process of workflow