A Novel Approach for Comparing Selected Metabolites in Citrus Leaves and Fruits Across Datasets

Abstract

Citrus fruits are valued not only for their nutritional benefits but also for their rich phytochemical content. Metabolomics has emerged as a comprehensive technique for assessing the chemical composition of fruits. The botanical connection between leaves, flowers, and fruits is reflected in both their structure and chemical composition, particularly in the flow of nutrients between plant organs. We introduced a new logarithm ratio-based approach to compare metabolite profiles between fruits and leaves. We hypothesize that this method allows for the analysis of multiple citrus metabolomic profiles to reveal known and novel correlation patterns, reflecting the dynamic connections between metabolic sources. To test this hypothesis, we leveraged comprehensive leaf metabolomic profiles from over 200 accessions in the Givaudan Citrus Variety Collection and reviewed published metabolomics data for fruits and juices of matching citrus types. By employing logarithm-transformed metabolic ratios within each dataset, we accounted for systematic differences across metabolomic platforms, achieving an unbiased analysis.

Keywords: citrus; comparison; datasets; metabolomics; method.

Figure 1

Citric-acid-to-malic-acid ratio in citrus fruit and leaves. This figure compares the citric-acid-to-malic-acid-ratio in the fruits and leaves of four mandarin varieties. Fruit data were sourced from Feng et al. (2018) [8], while leaf data were collected from the Givaudan Citrus Variety Collection at UC Riverside, representing similar but not identical mandarin varieties. Because each sample lacked replicates, error bars could not be calculated. The red dotted lines represent the standard error of the existing samples displayed on the chart.

Figure 2

Naringenin-to-hesperidin ratio comparison. This figure compares naringenin-to-hesperidin ratios between a leaf metabolomics dataset and two fruit datasets, calculated using Equation (1) (Section 2). Error bars are unique to each sample. Samples without error bars lacked sufficient replicates.

Figure 3

Proline-to-valine ratio comparison. This figure compares proline-to-valine ratios between a leaf metabolomics dataset and two fruit datasets, calculated using Equation (1) (Section 2). Error bars are unique to each sample. Samples without error bars lacked sufficient replicates.

Figure 4

Glutamate-to-glutamic-acid ratio comparison. This figure compares glutamate-to-glutamic-acid ratios between a leaf metabolomics dataset and two fruit datasets, calculated using Equation (1) (Section 2). Error bars are unique to each sample. Samples without error bars lacked sufficient replicates.

Figure 5

Volcano plot comparing metabolite ratios across multiple datasets (including fruit and leaf tissues) and citrus categories. The x-axis represents the mean log ratio of two selected metabolites, with negative values indicating a higher abundance of the denominator metabolite and positive values indicating a higher abundance of the numerator metabolite. The y-axis represents the significance level for the difference in the mean log ratio from 0, with a threshold set at -log₁₀(0.05) (one-sample t-test). (a) Examines the ratio between naringenin hesperitin. (b) Examines the ratio between phenylalanine and tryosine. (c) Examines the ratio between apigenin and luteolin. (d) Examines the ratio between arginine and proline. (e) Examines the ratio between glutamic acid and glutamine. (f) Examines the ratio between leucine and isoleucine.

Figure 6

(a) Volcano plot of untargeted metabolites (flavonoids): This plot identifies significant differences in flavonoid ratios between fruit and leaf datasets. The x-axis represents the difference in average log ratios between fruits and leaves, with positive values indicating higher log ratios in fruits and negative values indicating higher log ratios in leaves. The y-axis represents the significance level, with a threshold of -log₁₀(0.05). Each data point represents a specific flavonoid ratio. (b) Volcano plot of targeted metabolites (amino acids and common acids): This plot identifies significant differences in ratios of targeted metabolites (amino acids and common acids) between fruit and leaf datasets. The x-axis and y-axis interpretations are the same as in Figure 6a. See Equation (2) for the calculation [28].