Introduction

Potatoes (Solanum tuberosum L.) are currently the third most important crop worldwide after wheat and rice. It is also the most important vegetable-tuber crop for humans (Champouret 2010). Various diseases significantly reduce the quality and production of the potato crop. For example, late blight (LB) is caused by the oomycete Phytophthora infestans (Mizubuti and Fry 2006; Boevink et al. 2020; Ivanov et al. 2021). Different symptoms, such as wilting, death, and tuber rot, are induced by the LB disease on potatoes, resulting in destroying the potato yield in only a few days (Whisson et al. 2016). LB pathogen causes multi-billion dollar losses worldwide annually. LB is managed with fungicides, but the repetitive usage of fungicides has a negative influence on the environment and boosts the fast development of several effector loci in the pathogen genome (Raffaele et al. 2010). Discovering resistance (R) genes is considered the principal strategy for understanding resistance mechanisms to LB. At the genetic level, transcriptome sequencing is an effective tool for discovering R genes (Li et al. 2022). High-throughput DNA sequencing or Next-generation sequencing (NGS) or De Novo assembly methodology, is considered an important tool in genomics research. Thus, millions of DNA fragments can be sequenced at once using NGS, providing detailed data on genome structure, changes in gene behavior, genetic variants, and gene activity (Metzker 2010). Genome sequences of potato and P. infestans have been published (Xu et al. 2011), making sequencing-based transcriptome studies (RNA-seq) more accessible to potato LB studies. Up to now, RNA-seq research focused on potato-microbe interactions in the resistant and susceptible lines is little. Previous studies also performed ribonucleic acid-sequencing (RNA-seq) of potatoes infected by P. infestans pathogen to study alterations in the transcriptome. They carried out bioinformatics analyses to identify a large number of specific and wide-range transcription factors (TFs) and R loci that are associated with P. infestans infection (Stefańczyk et al. 2020). The first investigation of potato responses to pathogen attacks using RNA-seq (or transcriptome dynamics) was provided by Gao et al. (2013). Differentially expressed (DE) gene sets that discriminate resistant and susceptible lines were discovered, implying that Rpi-blb1 (RB) gene-induced tuber defense is mediated by powerful regulation of defense loci and ontology bins, such as hypersensitive-induced reaction protein and WRKY transcription factors. In the +RB transgenic line, a faster and higher expression of defense-related genes is associated with improved LB resistance. RNA-seq was used to build comprehensive transcriptome sets from two locally cultivated potato varieties that differed in their responses to LB infection.

The aim of this study was to observe and understand changes in the transcriptome between susceptible and resistant potato varieties to P. infestans and to identify DEG genes and ontology bins that discriminate compatible and incompatible interactions in susceptible and resistant varieties, respectively. In addition, a bioinformatics analysis was performed.

Materials and methods

Plant materials

Two locally cultivated potato varieties, the highly resistant Jelly (HR) and moderately susceptible Annabelle (MS), were selected from evaluation assays conducted over two consecutive seasons based on disease severity (DS) values (Mahfouze et al. 2021). Cultivars were grown under greenhouse conditions in pots (25 cm in diameter) filled with a mixture of sterilized sand and soil (1:1 v/v). A randomized complete block design with three replications was used, with five potatoes in their own pots per replication. Five tubers of each cultivar were used as a control.

Inoculation of potato varieties with P. infestans

Thirty days after planting, ten plants of each variety were sprayed with inoculum using an encysted zoospore suspension from P. infestans at a concentration of 5 × 104 sporangia ml−1 until the leaf surfaces were fully saturated (Chen et al. 2003). Inoculated potato plants were observed weekly post-inoculation for the development of LB symptoms. These varieties were evaluated in the greenhouse for two successive seasons: 2019/2020 and 2020/2021.

RNA isolation and transcriptome sequencing

The total RNA was extracted from two potato varieties inoculated with P. infestans and their controls using the TRIzol® Reagent (Invitrogen) via the manufacturer’s protocol, and they were treated with RNase-free DNaseI (Invitrogen). Nanodrop™ 2000 spectrophotometer, agarose gel electrophoresis, and Bioanalyzer 2100 (Agilent) were used to check and confirm the quantity and quality of RNAs. All RNAs were sent to the Beijing Genomic Institute (BGI) in China for library preparation and transcriptome sequencing using the IlluminaBGISEQ-500 platform to generate paired-end (2 × 150 bp) reads.

De novo transcriptome assembly and functional annotation

Clean data was extracted from raw data by deleting reads with low-quality (more than 20% of base qualities were less than 10), adaptors, and unknown bases (N bases greater than 5%). These were filtered to get clean reads and put in FASTQ format. TGICL (Pertea et al. 2003) was used to cluster transcripts to unigenes, and Trinity software (Grabherr et al. 2011) was used to do de novo assembly with clean reads (PCR duplication was deleted to enhance efficiency). The unigenes were separated into the following types: cluster and singleton. A cluster has the prefix CL and the cluster id behind it (in one cluster, there are numerous unigenes with more than 70% similarity), and a singleton has the prefix unigene. The assembled sequences were functionally annotated using NCBI protein (Nr) and nucleotide (Nt) databases, Swiss-Prot, Protein Family (PFAM), Gene Ontology (GO), EuKaryotic Orthologous Groups (KOG), and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases using the Basic Local Alignment Search Tool (BLASTX) a significance threshold of an e-value of 1 × 10−5.

Unigene functional annotation, unigene coding DNA sequences (CDS), and TF prediction

BLASTN and BLASTX (Altschul et al. 1990) were used to align unigenes to NT, NR, KOG, KEGG, and the SwissProt database to do the annotation. Blast2GO (Conesa et al. 2005) was used with NR annotation to do the GO annotation, and InterProScan5 (Quevillon et al. 2005) was used to do the InterPro annotation.

Transdecoder v3.0.1 was used to identify the candidate coding area. The longest open reading frame (ORF) was extracted, and the PFAM protein homologous sequences were searched by BLAST, SwissProt, and Hmmscan to predict the coding region. A TF is a protein that binds to specific deoxyribonucleic acid (DNA) sequences, thereby controlling the rate of transcription of genetic information from DNA to messenger RNA (mRNA). Getorf (Rice et al. 2000) was used to find the ORF of each unigene, after which the ORF was aligned to TF domains (PlntfDB) (http://plntfdb.bio.uni-potsdam.de/v3.0/) using the HMMER3 search (Mistry et al. 2013), and TFs were identified.

Unigene expression and time series analysis

Using Bowtie2 v2.2.5 (Langmead and Salzberg 2012), clean reads were mapped to unigenes, and RSEM was used to calculate the gene expression level (Li and Dewey 2011). The princomp function of R and hclust function were used to perform hierarchical clustering and PCA analyses, respectively. Clustering analysis is a common method of analyzing gene expression. The two kinds of clustering are hard and soft. The vast majority of clustering algorithms used produce hard partitions of the data, such as each gene or protein being assigned to exactly one cluster. Hard clustering is favorable if clusters are well separated. However, this is generally not the case for gene expression time-course data, where gene/protein clusters frequently overlap. Additionally, hard clustering algorithms are often extremely sensitive to noise. Mfuzz analysis software (Kumar and Futschik 2007) was used to do the soft cluster, as it is more suitable for gene expression data.

DEG detection and clustering analysis

DEGs were detected with PossionDis (Fold Change ≥2.00 and False discovery rate (FDR) ≤0.001), which is based on Poisson distribution described by Audic and Claverie (1997). Hierarchical clustering for DEGs was carried out using Pheatmap. For clustering with more than two groups of DEGs, we performed the intersection and union DEGs between them, respectively.

Gene ontology analysis and DEG pathway analysis

DEGs were categorized based on the GO annotation result. With the KEGG annotation result, DEGs were categorized according to the official classification and pathway functional enrichment, as well as GO functional enrichment, and were carried out using phyper, a function of R. The p-value calculating formula in hypergeometric test was as follows:

$${\text{P}} = 1 - \sum\limits_{i = 0}^{m - 1} {\frac{{\left( \eqalign{M \hfill \cr i \hfill } \right)\left( \eqalign{ N - M \hfill \cr n - i \hfill } \right)}}{{\left( \eqalign{ N \hfill \cr n \hfill } \right)}}}$$

FDR was calculated for each p-value. In general, the terms with FDR no larger than 0.01 were defined as significant enrichment (https://en.wikipedia.org/wiki/Hypergeometric_distribution).

Detection of plant disease resistance genes

To detect potential disease R genes by evaluating the query coverage and identity values, a BLAST analysis was carried out by mapping the unigene to the Pathogens–related genes (PRG) (Sanseverino et al. 2013) database v2.0 (http://prgdb.crg.eu/).

Results

De novo assembly and sequencing reads filtering

Adaptor-polluted, low-quality, and high-content unknown base (N) reads were eliminated from the sequencing sequences before the downstream analysis. The original data and clean reads quality metrics are shown in Table 1. The raw data presented in FASTQ format was recorded in the NCBI Sequence Read Archive (SRA) under accession number PRJNA722040, resulting in a total of 17.77 Gb. In HR, the base quality score Q30 was 91.74%, compared with the control’s score of 92.01%. However, the quality score Q30 of MS was 91.65%, compared with the control’s score of 91.43%. This indicates that the four potato samples had an acceptable sequencing quality. Trinity software was used to assemble the data and overlapping information in high-quality reads.

Table 1 Clean reads quality metrics of the two potato cultivars; Jelly and Annabelle and their controls
Full size table

De novo assembly

To create a reference sequence for further analysis in the project, clean reads were assembled following sequencing. Trinity software was used to carry out de novo assembly with clean reads following filtering. The quality metrics of transcripts and unigenes of HR and MS compared with the control shown in Table 2. Clustering yielded a total of 382,255 transcripts for the four potato samples, with a total length of 2517 bp in the assembled part. Further assembly generated 96,255 unigenes, with a total length of 78,894,974 bp, and an average length of 819 bp. The N50 and GC content of unigenes were 1336 bp and 39.82%, respectively (Table 2).

Table 2 Quality metrics of transcripts and unigenes of the four potato samples
Full size table

Unigene functional annotation

After assembly, the unigenes were functionally annotated by aligning with seven different functional databases (GO, KOG, NR, KEGG, NT, SwissProt, and InterPro). Annotated unigenes aligned with seven functional databases revealed 52,237 (NR: 54.27%), 57,829 (NT: 60.08%), 35,894 (Swiss Prot: 37.29%), 42,663 (KOG: 44.32%), 38,948 (KEGG: 40.46%), 36,109 (GO: 37.51%), and 46,211 (InterPro: 48.01%) as are listed in Table 3.

Table 3 Annotation summary
Full size table

Unigene coding sequence (CDS) and TF prediction

The candidate coding region among unigenes was identified using TransDecoder software. For the CDS prediction and longest ORF were chosen and blasted using SwissProt and Hmmscan databases to look for PFAM protein homologous sequences. A total of 42,602 ORFs were found, and the total length of the CDS was 38,177,808. The maximum length of CDS was 8418, while the minimum length was 297. The G and C bases in all CDS were 43.28%. There were a total of 36,018 complete ORFs, and the length distribution was analyzed. After assembly, unigenes that encode TF were predicted (Table S1), and classification on the TF family was conducted. In the current study, a total of 1874 putative TFs were assigned and grouped into 85 families (Table S2). The TFs found in the four S. tuberosum samples play role in the plant defense system. In this study, TFs belong to the MYB (193, 10.30%), AP2-EREBP (149, 7.95%), MYB-related (147, 7.84%), C2H2 (136, 7.26%), WRKY (113, 6.03%), bHLH (110, 5.87%), and C3H (107, 5.71%) families.

The distribution of unigene expression level

Depending on the level of gene expression, a box plot was adopted to demonstrate the distribution of the gene expression level (Fig. 1). To represent the gene expression in HR, MS, and the control, each unigene was analyzed more intuitively using fragment per Kb million (FPKM) intervals. Figure 1 shows the gene amount in different FPKM intervals. The highest gene expression was obtained from the MS control, with 13,410 FPKM values greater than 10, followed by the HR control with 12,845 and MS with 12,841, while HR showed lower FPKM values with 12,329. However, FPKM ≤ 1 means with an extremely low expression level with FPKM values less than 10 were obtained from HR at 23,534, followed by the MS control with 21,662 and MS with 17,708 (Fig. 1).

Fig. 1
figure 1

a Box plot represent the gene expression of the two potato varieties; Jelly highly resistant to P. infestans, and Annabelle the moderately susceptible as well as their controls. b Plot of expression level distribution of unigenes of the four potato samples in fragment per kb million (FPKM) intervals. X-axis refers to the sample name and Y-axis refers to the number of genes. The depth of the color represents different gene expression levels: FPKM ≤ 1 means extremely low expression level, FPKM is between 1 and 10 means genes have medium expression levels, and FPKM ≥ 10 are genes with high expression levels. c Venn diagram analysis of the four potato samples. d Up-regulated and down-regulated of the two groups (HR and MS); Jelly cont. vs. resistant (HR) and Annabelle cont. vs. susceptible (MS) DEGs. X axis represents comparison functionality between each group. Y axis represents DEG numbers. Red color acts up-regulated DEGs. Blue color acts down-regulated DEGs

Full size image

Gene expression level between potato varieties resistant and susceptible to LB

The expression level data of the two potato varieties showed that the amount of specific expression genes was 7263 and 17,572 in HR and the control, respectively. There were 60,218 genes shared between them. However, 11,798 and 6907 deferential expression genes (DEGs) were found to be unique in MS and the control, respectively. There were 67,580 genes shared by both (Fig. 1).

DEG detection and clustering analysis of DEG

Depending on the level of gene expression, DEGs between HR and MS were identified and detected using DEGseq, DEseq2, EBseq, NOIseq, and PossionDis algorithms. The two groups were compared as follows: HR control versus HR resistant and MS control versus susceptible MS were further analyzed among the 65,535 total DEGs numbers. The low percentage of KEGG annotations of HR occurred on upregulated DEGs with 30,565 versus 47,340 of MS, while downregulated DEGs were higher in HR with 34,970 versus 18,159 in MS (Tables S3 and S4). A total of 19,538 DEGs (upregulated: 5909; downregulated: 13,629) were scored in HR, and 10,756 DEFs (upregulated: 7408; downregulated: 3348) were scored in MS. Consequently, the upregulated DEGs in MS were higher compared to HR (Fig. 1). Therefore, the MA plot, Scatter plot, Volcano plot, and Heatmap plot were used to show the distribution of DEGs in HR and MS (Figs. 2 and 3). Additionally, the hierarchical clustering was carried out for data analysis for DEGs via synthesis clusters of genes with identical patterns of expression (Fig. 4).

Fig. 2
figure 2

a MA plot of DEGs. X axis refers to value A (log2 transformed mean expression level) and Y axis refers to value M (log2 transformed fold change). b Scatter plot of DEGs. X Y axis represents log10 transformed gene expression level. c Volcano plot of DEGs. X axis represents −log10 transformed significance and Y axis represents log2 transformed fold change. Red points: up regulated DEG, blue points: down regulated DEG. black points: non-DEGs

Full size image
Fig. 3
figure 3

Heatmap of DEGs. Y axis refers to DEGs. The dark color means the high expression level while; the light color means the low expression level

Full size image
Fig. 4
figure 4

Heatmap of Hierarchical clustering of DEGs. X axis represents the comparisons for clustering analysis. Y axis represents DEGs. Coloring indicate fold change (high: red, low: blue)

Full size image

Gene ontology analysis, pathway, and functional annotation of DEGs

The GO analysis (http://www.geneontology.org/) classified 46,248 unigenes in the HR control versus the resistant HR and 26,921 unigenes in the MS control versus susceptible MS into the following three categories: biological process, cellular component, and molecular functions (Fig. 5). In the cellular component category, most genes in HR and MS were classified into seven subcategories, and the number of genes in all subcategories was higher in HR compared to MS. The two groups were compared as follows: cells with 4342 genes in HR versus 2576 in MS, respectively (Fig. 5). In the biological process category, all subcategories were higher in HR compared to MS. In comparing HR to MS among the top subcategories, the cellular process scored 3899 genes in HR versus 2457 in MS, the metabolic process scored 3821 versus 2421 unigenes, the biological regulation scored 1234 versus 569, and the regulation of biological process scored 1053 versus 466, and the biogenesis scored 737 in HR versus 494 in MS. In the molecular functions category, the highest two subcategories were catalytic activity with 4578 in HR versus 2652 unigenes in MS and binding with 4461 in HR versus 2478 unigenes in MS. All other subcategories were higher in HR compared to MS. Figure 6 shows the GO classification of upregulated and downregulated genes in the HR control versus the resistant HR and MS control versus susceptible MS. The GO enrichment result was displayed using a directed acyclic graph with each node representing the name of the GO database and the p-value. The lower p-value points to more significant enrichment (Fig. 7).

Fig. 5
figure 5

GO classification of DEGs. X axis: number of DEG. Y axis: GO term

Full size image
Fig. 6
figure 6

GO classification of up regulate and down regulate genes. X axis: GO term. Y axis: the amount of up/down regulate genes

Full size image
Fig. 7
figure 7

GO functional enrichment of DEGs

Full size image

Metabolic pathways of unigenes were studied in two groups (HR and MS) to recognize the specific processes, gene functions, and gene interactions using the KEGG. Pathway enrichment was further analyzed to get more insights into classification and functional enhancement in HR and MS potato varieties. A total of 10,100 unigenes in the HR control versus resistant HR and 6330 MS control versus susceptible MS were KEGG-annotated unigenes and enriched for 20 subclasses of five different functional pathways (cellular processes, environmental information processing, genetic information processing, metabolism, and organismal systems) (Fig. 8). HR and MS groups were compared as follows: 5901 genes versus 3810 were classified into the category “metabolism,” predominantly involved in “global and overview maps” (2356 versus 1527), and “amino acid metabolism” (517 versus 304). Among “genetic information processing,” the majority of the unigenes were involved in “translation” (1129 versus 893), and “transcription” (564 versus 362). On the other hand, “Biochemical pathways”, 38,948 DEGs were identified in the HR versus the resistant HR and MS control versus susceptible MS, whereas 10,397 HR and 6120 for MS were mapped to 137 KEGG pathways (Fig. 8). Most genes were annotated with a metabolic function (38,948 unigenes), followed by “genetic information processing” (4033 unigenes), and “cellular processes” (799 unigenes). The top number of unigene pathways represented in the “metabolism category” in HR versus MS were metabolic pathways (ko01100; 2165, 20.82% versus 1416, 23.14%), biosynthesis of secondary metabolites (ko01110; 1168, 11.23% versus 736, 12.03%), biosynthesis of amino acids (ko01230; 247, 2.38% versus 158, 2.58%), and purine metabolism (ko00230; 239, 2.3% versus 170, 2.78%). The two pathways in the “organismal systems” (environmental adaptation) category were plant-pathogen interaction (ko04626; 374, 3.6% versus 144, 2.35%) and circadian rhythm-plant (ko04712; 161, 1.55% versus 106, 1.73%). The complete set of matched pathway functional enrichment results of the HR control versus resistant HR and MS versus susceptible MS are presented in Tables S5 and S6.

Fig. 8
figure 8

Pathway classification of DEGs using functional classification of KEGG. X axis: number of DEG. Y axis: functional classification of KEGG

Full size image

Figure 9 shows the 11 following pathways that were significantly enriched in HR: anthocyanin biosynthesis, benzoxazinoid biosynthesis, flavonoid biosynthesis, stilbenoid, diarylheptanoid and gingerol biosynthesis, sesquiterpenoid and triterpenoid biosynthesis, selenocompound metabolism, vitamin b6 metabolism, photosynthesis, brassinosteroid biosynthesis, and isoflavonoid biosynthesis. However, another 12 pathway functions were significantly enriched in MS as follows: glyoxylate and dicarboxylate metabolism, arachidonic acid metabolism, biosynthesis of unsaturated fatty acids, one carbon pool by folate, pyrimidine metabolism, purine metabolism, oxidative phosphorylation, nitrogen metabolism, phagosome, fatty acid elongation, fatty acid metabolism, and citrate cycle (TCA cycle). The complete set of matched pathway functional enrichment results of the HR control versus resistant HR and MS control versus susceptible MS are presented in Tables S5 and S6.

Fig. 9
figure 9

Pathway functional enrichment of DEGs. X axis represents enrichment factor. Y axis represents pathway name. The color indicates the q value (high: white, low: blue), the lower q value indicates the more significant enrichment. Point size indicate DEG number (The bigger dots refer to larger amount). Rich factor refers to the value of enrichment factor, which is the quotient of foreground value (the number of DEGs) and background value (total gene amount). The larger the value, the more significant enrichment

Full size image

Detection of pathogen-resistance genes (R)

The results found that 2407 DEGs were R genes, of which known loci were annotated in detail. These annotated loci involved RLP type, TNL type, CNL type, RLK-GNK2 type, N type, RLK-Malectin, NL type, TN type, TNL-OT type, CNL-R1, TNL-TNL type, CN type, PTO-like type, CNL-R1, MLO-like type, L type, T type, TL type, CNL-RPW8 type, and other types. These candidate genes can be used to detect potato varieties containing disease R genes (Table S7).

Discussion

In the current investigation, RNA-seq was used to obtain a high-quality transcriptome of highly resistant (HR), moderately susceptible (MS), and their controls to P. infestans infection. A few reports have been published on the S. tuberosum RNA-seq transcriptome with LB resistance. A large number of transcript sequences were included in several biological processes in this study. The sequence information from the present study is publicly available to aid in facilitating a better understanding of LB disease resistance for future potato improvement. As mentioned in the PGSC (2011), cultivated potatoes have a narrow genetic base, resulting originally from limited germplasm. Furthermore, most potato varieties are autotetraploid, highly heterozygous, suffer from severe inbreeding depression, and are susceptible to multiple pathogens. However, the varieties used in the study were neither transgenic lines nor hybrids with wild Solanum species to transfer specific LB resistance genes. They were well characterized by evaluation assays conducted over two consecutive seasons based on the DS values. One was highly resistant (HR) and the other was moderately susceptible (MS) to P. infestans infection. Therefore, this is the first report to our knowledge that used RNA sequencing-based transcriptome analyses of these varieties. The results provide novel data concerning the molecular mechanism of potato resistance to P. infestans disease. Wang et al. (2009) mentioned that RNA-seq analysis is a crucial and cost-effective method for analyzing transcriptome data. Plant responses to pathogenic diseases have been studied using RNA-seq technology for a long time.

In this investigation, the base quality score Q30 in HR, was 91.74%, compared with the control’s score of 92.01%. However, the quality score Q30 of MS was 91.65%, compared with the control’s score of 91.43%. This indicates that the four potato samples had an acceptable sequencing quality. Similar results were observed by Gao et al. (2013) extracted RNA of 36 samples from potato plants infected with LB and water-inoculated ones and were subjected to RNA-seq. About 483 million paired end reads were produced, yielding a mean of 13.4 million reads per sample. Quality filtering was successful in 436.3 million readings (90.3%). In this work, a total of 19,538 DEGs (upregulated: 5909; downregulated: 13,629) were recorded in HR, and 10,756 DEFs (upregulated: 7408; downregulated: 3348) were scored in MS. Consequently, the upregulated DEGs in MS were higher compared to HR. Similar results were recorded by Li et al. (2023) indicated that the number of DEGs in the grape variety susceptible VvMF to white rot disease was higher than that in the resistant variety Vd0940, due to significant variations in response mechanisms between the two grape varieties after infection by Coniella diplodiella (Speg.).

In this finding, RNA sequencing, a comparative DEG analysis, evaluated the transcriptome expression responses of HR and MS to P. infestans. Results indicated that many genes were upregulated in HR compared with MS. These involved biosynthesis of secondary metabolites, carbon metabolism, biosynthesis of amino acids, purine metabolism, plant-pathogen interaction and circadian rhythm-plant, cellular processes (endocytosis, phagosome, and peroxisome), genetic information processing (spliceosome, ribosome biogenesis in eukaryotes, and protein processing in the endoplasmic reticulum), MAPK, plant hormone signal transduction, and phosphatidylinositol signaling system (Nakashima and Yamaguchi-Shinozaki 2013; Ren and Gray 2015; Verma et al. 2016). On the other hand, about 2407 DEGs belonging to R genes were indicated. Among them are known genes, such as PTO-like and N, but most of them are unknown novel genes, which will supply a large number of gene candidates to detect potato varieties resistance to P. infestans. They may play an important role in the resistance of the host against P. infestans through the production of toxic activities by these metabolites (He et al. 2021). Qian et al. (2017) found that plants have defense mechanisms that depend on a molecular immunity system through R genes that permit them to resist different pathogens involving fungi, oomycetes, bacteria, and viruses. Most R genes encode nucleotide binding sites (NBS) and composed of three domains that permit them to be divided into TIR-NBS-LRR (TNL), CC-NBS-LRR (CNL), or RPW8-NBSLRR (RNL), of which the C-terminal LRR domain exhibits high diversity and has been linked to pathogen recognition. At the beginning of the infection with LB pathogen, underexpressed genes increased, especially NBS-LRR (nucleotide binding site, leucine-rich repeats), which could indicate that at the beginning, several underexpressed genes are needed to produce resistance in the late stages of infection (Yang et al. 2018).

In this regard, KEGG enrichment analysis of DEGs in HR variety revealed 11 pathways, including anthocyanin biosynthesis, benzoxazinoid biosynthesis, flavonoid biosynthesis, stilbenoid, diarylheptanoid and gingerol biosynthesis, sesquiterpenoid and triterpenoid biosynthesis, selenocompound metabolism, vitamin b6 metabolism, photosynthesis, brassinosteroid biosynthesis, isoflavonoid biosynthesis, and phenylpropanoid biosynthesis. The latter is necessary for plants because it plays a main role in the production of hydroxycinnamyl alcohol, which serves as the building block of lignin and refers to structure support and resistance against LB pathogen (Yogendra et al. 2015). However, MS variety showed 12 pathway functions involving glyoxylate and dicarboxylate metabolism, arachidonic acid metabolism, biosynthesis of unsaturated fatty acids, one carbon pool by folate, pyrimidine metabolism, purine metabolism, oxidative phosphorylation, nitrogen metabolism, phagosome, fatty acid elongation, fatty acid metabolism, and citrate cycle (TCA cycle).These pathways play an important role in the defense response to pathogen (Chen et al. 2016; Liu et al. 2022).

Conclusion

The first RNA-seq (or transcriptome dynamics) was performed on HR and MS to the P. infestans infection and their controls. Up to now, no data about transcriptome changes in response to these cultivars to LB was available. RNA-seq analysis identified P. infestans-responsive genes and their expression in potatoes. We investigated the mechanism of the response of these cultivars to the P. infestans pathogen via TFs and R genes, which play an important role in the defense response. These genes were identified via transcriptome analysis during infection with P. infestans. Besides, KEGG and GO analysis of the DEGs provided some insights into molecular mechanisms underlying LB pathogen resistance in HR and MS. This work will contribute in the genetic improvement of the tolerance to LB disease in potato.