Weighted Gene Co-Expression Network Analysis Uncovers Core Drought Responsive Genes in Pecan (Carya illinoinensis)

Abstract

Drought severely affects the growth and production of pecan (Carya illinoinensis), while genes conferred drought adaptation are yet to be fully elucidated. Here, an in-depth exploration of the two different RNA-seq projects regarding drought stress (designated as P1 and P2) was performed via weighted gene co-expression network analysis. For the two projects, there existed one pair of modules (P1 turquoise module and P2 blue module) that was probably associated with drought resistance, as the paired modules both exhibited an increased expression profile with increasing water shortage stress and were annotated to be involved in oxidative stress response and the signaling pathways of abscisic acid and jasmonic acid. There were 441 and 1258 hub genes in the P1 turquoise module and P2 blue module, respectively, among which, 140 were overlapped and thus were recognized as core drought responsive genes. An additional drought stress experiment was conducted for RT-qPCR validation, and the results showed that the 20 core genes selected for detection were highly responsive to water deficit. Together, our results will be helpful for understanding the molecular mechanism of drought response and improving drought resistance in pecan.

Keywords: core genes; drought; pecan; weighted gene co-expression network analysis.

Figure 1

The number and expression patterns of differentially expressed genes (DEGs) for pecan in response to drought stress derived from two different projects (P1 and P2). (a) Experimental scheme depicting the drought treatment in P1 and P2. P1 includes the dehydration (D) and rehydration (RD) periods, and P2 contains only water-deficit treatment (T). (b) The number of DEGs for P1 and P2. (c) Expression patterns for the 4592 common DEGs are grouped into six clusters (C1–C6). Expression levels for each project are independently z-score normalized.

Figure 2

Hierarchical clustering of the genes used for co-expression network construction. (a) Genes clustering in project 1 (P1). (b) Genes clustering in P2. For the clustering, y-axis suggests expression dissimilarity between neighboring genes, and each branch indicates one gene. Module colors represent the module designation for the co-expressed genes.

Figure 3

Expression pattern and gene ontology functional enrichment for the modules in project 1. Gene expressions are z-score normalized and visualized via line charts as well as heatmaps. D0–D24 represent the samples collected 0~24 days after dehydration. RD is the sample collected after rehydration. Then, -1, -2, and -3 are the three different biological replications.

Figure 4

Expression pattern and gene ontology functional enrichment for the modules in project 2. Gene expressions are z-score normalized and visualized via line charts as well as heatmaps. CK–T15 represent the samples collected 0~15 days after drought stress. Then, -1, -2, and -3 are the three different biological replications.

Figure 5

The number of hub genes (a) and the expression patterns for the common genes (b) in the turquoise module of project 1 (P1) and the blue module of P2. Expression patterns are z-scores independently for P1 and P2.

Figure 6

The expression patterns of 20 core-drought responsive genes detected via RT-qPCR. The detached leaves were sampled after 0, 6, 12, and 24 h of water loss.