*3.3. Transcriptome Data Quality Analysis*

We constructed nine sequencing libraries with three replicates of each treatment. The number of raw reads per sample ranged from 45.57 to 47.33 million, as shown in Table S2. Clean reads exceeded 42.33 million for all samples except for some low-quality reads, splices and fuzzy nucleotides. The number of clear reads obtained from the cDNA libraries of each experiment indicates that the gene abundance and transcript length are sufficient. In our libraries, the percentage of clean reads was not less than 90%, with Q20 and Q30 values exceeding 96.66% and 88.09% for all samples, respectively. This indicates that the data quality of the transcriptome is sufficient (Table S2).

#### *3.4. Metabolomics Assay*

The type and content of metabolites change under different stimuli leading to phenotypic changes. To investigate the effect of exogenous spermidine on changes in the lettuce metabolome under heat stress, we analyzed the major and minor metabolites in leaves. We quantified the relative levels of all filtered ions and performed PCA analysis (Figure 3B). As shown in the figure, the first two principal components explained 48% of the total variation, indicating a clear separation between water and spermine treatments under heat stress, suggesting that the identified metabolites play an important role in mitigating heat tolerance in lettuce under the influence of exogenous spermine(Figure 3B).Then, PLS-DA models (H vs. CK, HS vs. CK, HS vs. H) between each two groups were established using metaX software to screen out different metabolites, and when the values of model parameters (R2 and Q2) were high, it indicated that the current PLS-DA model was more reliable(Table 3). To further identify potential metabolites for water treatment and spermine treatment under high temperature stress, "candidate" metabolites with the following characteristics were included in the analysis with the screening conditions: (1) VIP ≥ 1; (2) fold change ≥ 1.2 or

≤ 0.8333; (3) *p*-value < 0.05, and the three were taken to intersect to obtain the common metabolite as the differential metabolite.

**Figure 3.** Differential metabolome analysis in HSvsH. (**A**) Differential metabolite composition. (**B**) PCA model for metabolic data. (**C**) Volcano plot of differential metabolites. (**D**) Heat map for cluster analysis of differential metabolites. CK: 22 ◦C/17 ◦C, distilled water; H: 35 ◦C/30 ◦C, distilled water; HS: 35 ◦C/30 ◦C, distilled 1 mM Spd.

**Table 3.** PLS-DA model parameters.


CK: 22 ◦C/17 ◦C, distilled water; H: 35 ◦C/30 ◦C, distilled water; HS: 35 ◦C/30 ◦C, distilled 1 mM Spd.

Compared with CK, 897 metabolites were identified under high temperature treatment and could be classified into 14 classes, while 1052 differential metabolites were identified in HSvsCK and classified into 15 classes; we obtained a total of 307 differential metabolites in HSvsH Supplementary Table S3). These metabolites included 7 alkaloids and derivatives, 75 benzenoids, 3 hydrocarbon derivatives, 60 lipids and lipid-like molecules, 8 nucleosides, nucleotides and analogues, 56 organic acids and derivatives, 7 organic nitrogen compounds, 26 organic oxygen compounds, 49 organic heterocyclic compounds and 16 phenylpropanoids and polyketides (Figure 3A).

In the comparison of the different treatments, "organic acids and their derivatives", "organic heterocyclic compounds", "lipids and lipid-like molecules" and "benzenes" were noted. Similarly, we found a number of "phenyl propane and polyketide compounds" in HS vs. H, and the phenyl propane pathway is the main pathway for the synthesis of flavonoids. It is noteworthy that we also found KEGG enrichment in the transcriptome for the flavonoid biosynthesis pathways, and the relative expression of most compounds was higher in the high-temperature applied spermidine treatment than in the high-temperature sprayed water treatment, as shown by clustering analysis and volcano plots (Figure 3C,D). This may further suggest that exogenous spermidine enhances the heat tolerance of lettuce by affecting flavonoid biosynthesis.

### *3.5. Analysis of DEGs*

The number of DEG groups among the three treatments are shown, and overlapping parts indicate the intersection of different combinations (Figure 4B). There were 2101 genes upregulated and 960 genes downregulated in lettuce leaves under high-temperature stress compared to the control (Figure 4C). Comparing the treatment with spermidine spray to distilled water, 818 genes were upregulated, and 284 genes were downregulated. In addition, a total of 2489 genes were upregulated and 717 genes were downregulated in the high-temperature treatment with spermidine compared to the control. However, when comparing among the groups the overlapping relationships of the differentially expressed genes, there were 3061 genes, 1102 genes and 2306 DEGs expressed in the comparison of HvCK, HS vs. CK and HS vs. H, respectively, while 151 genes were expressed in common among the three treatment groups. These results suggest that spermidine affects the response of plant genes to high-temperature stress to some extent, leading to changes in the number and type of DEGs.

**Figure 4.** Analysis of DEGs in lettuce under high-temperature stress with exogenous spermidine. (**A**) Transcriptome sample correlation map. (**B**) Venn diagrams of DEGs. (**C**) Numbers of DEGs. CK: 22 ◦C/17 ◦C, distilled water; H: 35 ◦C/30 ◦C, distilled water; HS: 35 ◦C/30 ◦C, distilled 1 mM Spd.
