Identification of biological pathway and process regulators using sparse partial least squares and triple-gene mutual interaction

Abstract

Identification of biological process- and pathway-specific regulators is essential for advancing our understanding of regulation and formation of various phenotypic and complex traits. In this study, we applied two methods, triple-gene mutual interaction (TGMI) and Sparse Partial Least Squares (SPLS), to identify the regulators of multiple metabolic pathways in Arabidopsis thaliana and Populus trichocarpa using high-throughput gene expression data. We analyzed four pathways: (1) lignin biosynthesis pathway in A. thaliana and P. trichocarpa; (2) flavanones, flavonol and anthocyannin biosynthesis in A. thaliana; (3) light reaction pathway and Calvin cycle in A. thaliana. (4) light reaction pathway alone in A. thaliana. The efficiencies of two methods were evaluated by examining the positive known regulators captured, the receiver operating characteristic (ROC) curves and the area under ROC curves (AUROC). Our results showed that TGMI is in general more efficient than SPLS in identifying true pathway regulators and ranks them to the top of candidate regulatory gene lists, but the two methods are to some degree complementary because they could identify some different pathway regulators. This study identified many regulators that potentially regulate the above pathways in plants and are valuable for genetic engineering of these pathways.

Transcriptional regulation of proanthocyanidin biosynthesis pathway genes and transcription factors in Indigofera stachyodes Lindl. roots

“…TGMI has been used to study lignin biosynthesis pathway in Arabidopsis thaliana [ 62 ], Populus [ 63 ], and Populus trichocarpa [ 64 ], for identifying which regulatory genes potentially control wood formation. In this study, we also applied the TGMI algorithm to true pathway regulators of PA biosynthesis in I. stachyodes roots based on the tissue-specific I. stachyodes gene expression datasets.…”

Integrative analysis of metabolome, proteome, and transcriptome for identifying genes influencing total lignin content in Populus trichocarpa

“…( Rao et al., 2019 ) and Hong et al. ( Hong et al., 2021 ) applied gene co-expression network analysis methods to explore genes and transcription factors related to the monolignol biosynthetic pathway in P. trichocarpa . However, their studies focused on establishing network relationships between genes without specifically analysing the impact of genes on lignin content.…”