NetMiner-an ensemble pipeline for building genome-wide and high-quality gene co-expression network using massive-scale RNA-seq samples

• Published in: PloS one Vol. 13; no. 2; p. e0192613
• Main Authors: Yu, Hua, Jiao, Bingke, Lu, Lu, Wang, Pengfei, Chen, Shuangcheng, Liang, Chengzhi, Liu, Wei
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 09.02.2018; Public Library of Science (PLoS)
• Subjects: Agronomy; Algorithms; Bioinformatics; Biological activity; Biology and Life Sciences; Case studies; Circular RNA; Computer and Information Sciences; Datasets; Developmental biology; Gene expression; Gene sequencing; Genes; Genetics; Genomes; Genomics; Laboratories; Methods; Non-coding RNA; Pharmacy; Phenotypes; Pipeline design; Plants (botany); Research and Analysis Methods; Ribonucleic acid; RNA; Transcription factors; Beijing China; China
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0192613

Accurately reconstructing gene co-expression network is of great importance for uncovering the genetic architecture underlying complex and various phenotypes. The recent availability of high-throughput RNA-seq sequencing has made genome-wide detecting and quantifying of the novel, rare and low-abundance transcripts practical. However, its potential merits in reconstructing gene co-expression network have still not been well explored. Using massive-scale RNA-seq samples, we have designed an ensemble pipeline, called NetMiner, for building genome-scale and high-quality Gene Co-expression Network (GCN) by integrating three frequently used inference algorithms. We constructed a RNA-seq-based GCN in one species of monocot rice. The quality of network obtained by our method was verified and evaluated by the curated gene functional association data sets, which obviously outperformed each single method. In addition, the powerful capability of network for associating genes with functions and agronomic traits was shown by enrichment analysis and case studies. In particular, we demonstrated the potential value of our proposed method to predict the biological roles of unknown protein-coding genes, long non-coding RNA (lncRNA) genes and circular RNA (circRNA) genes. Our results provided a valuable and highly reliable data source to select key candidate genes for subsequent experimental validation. To facilitate identification of novel genes regulating important biological processes and phenotypes in other plants or animals, we have published the source code of NetMiner, making it freely available at https://github.com/czllab/NetMiner.