Comprehensive meta-analysis and co-expression network analysis identify candidate genes for salt stress response in Arabidopsis

Abstract

Salinity is one of the major environmental limitations to agricultural productivity throughout the world. To overcome the defects of microarray sequencing analyses caused by heterogeneity among individual studies, the effect size (ES) analysis based on the NetworkAnalyst web tool was conducted to synthesize 13 salt stress-related microarray data-sets, filtered from over 2000 data-sets and publications. A total of 3811 differently expressed (DE) genes were identified with 1476 up-regulated and 2335 down-regulated genes in NaCl-treated samples. A total of 172 biological processes were enriched by the over-expressed genes, like “response to jasmonic acid”, “positive regulation of transcription, DNA-templated”, and “insulin resistance”. The co-expression network construction yielded 15,630 nodes and 73,125 edges, further clustered into 178 co-expressed clusters based on the Heuristic Cluster Chiseling Algorithm (HCCA) method. A total of 1036 genes with the top 5% of clusters were treated as hub genes. The identification of DE genes, gene-gene co-expressing relations, and hub genes may contribute to uncover the mechanisms of salinity stress tolerance in Arabidopsis.

Keywords:

• Arabidopsis thaliana
• salt stress
• meta-analysis
• co-expression network
• NetworkAnalyst

Abbreviations:

• DE – differently expressed
• TFs – transcriptional factors
• TRs – transcriptional regulators
• PKs – protein kinases
• HCCA – heuristic Cluster Chiseling Algorithm
• ABA – abscisic acid
• BRs – brassinosteroids
• JA – jasmonic acid
• CK – cytokinins
• DEGs – differently expressed genes
• ES – effect sizes
• FEM – fixed effects model
• REM – random effects model
• MCL – Markov Clustering
• FDR – false discovery rate
• GO – gene ontology
• HRR – highest reciprocal rank
• PCA – principal component analysis
• BP – biological processes
• MF – molecular functions
• MAPK – mitogen-activated protein kinase