Standardization of Transcriptome Analysis Improves Gene Selection for Sugarcane (Saccharum spp.) Breeding

Climate changes impose the urgency for the development of more drought tolerant sugarcane (Saccharum spp.), a crop with great importance. The identification of molecular mechanisms associated with improved tolerance to water deficit may accelerate sugarcane breeding by either transgenic or genomic selection approaches. Here, we investigated if the standardization of massive gene expression analysis may assist in determining potential candidate genes with higher precision. We analyzed 16 RNA-seq libraries publicly available at the Sequence Read Archive (SRA) from NCBI involving sugarcane and water deficit treatment as the common factors. All reads were mapped against the COMPGG transcriptome, a de novo assembly from six sugarcane cultivars, by STAR 2.7.10b and count tables were generated by SamTools 1.15.1. The count tables were then analyzed by the DESeq2 package for R, defining as differentially expressed genes (DEGs) those with p-values < 0.05 and log-fold change values > 2 for overexpressed genes and < -2 for underexpressed. DEGs were then ran using NetworkX (Python) into the biological network developed and published by Rody et al. (2021), which includes data mining of metabolic pathways from BioGRID and KEGG, and protein-protein interactions from SwissProt. The complete network was analyzed through Cytoscape 3.8.1, providing the complete annotation for every DEG. The set of most frequent DEGs between all the 16 libraries were submitted to a gene set enrichment analysis (GSEA), revealing 788 terms for biological processes (cellular process as the most frequent), followed by 252 terms for molecular function (protein binding as the most frequent), and 79 for cellular components (plasma membrane as the most frequent). Curiously, the most common gene among libraries was a Sucrose Galactosyltransferase, a component of the galactose metabolism pathway, responsible for the initial steps of the biosynthesis of galactose, glucose and fructose, being this whole process overrepresented with genes from the current analysis. Within all the 86 enriched pathways, the amino sugar and nucleotide sugar metabolism was the most represented with 29 genes. Therefore, we conclude that the standardization of the analysis from previously published data can provide precise information for gene prospection for functional analysis.