ncRNA_net: Development of a novel approach to lncRNA-mRNA Regulatory Network Inference
Main PI:Peter Kindgren, Department of Forest Genetics and Plant Physiology, Umeå Plant Science Centre (UPSC), Swedish University of Agricultural Sciences
Co-PI:Jian-Feng Mao, Department of Plant Physiology, Umeå Plant Science Centre (UPSC), Umeå University
Abstract
This research addresses a critical gap in plant biology: while thousands of long non-coding RNAs (lncRNAs) can be identified in plant species, their functional characterization remains severely limited. Unlike protein-coding genes with known regulatory mechanisms, lncRNA-mRNA relationships lack predictive frameworks, hindering hypothesis generation and functional studies.
Objectives
1. Develop ncRNA_net, a multi-tier computational framework integrating signed distance correlation, dual tree ensemble learning, graphical lasso, and dynamic inference components.
2. Reconstruct comprehensive signed lncRNA-mRNA regulatory networks from expression data to distinguish activation from repression.
3. Implement causal directionality inference using time-series and pseudotime data from bulk and single-cell experiments.
4. Validate the framework through extensive benchmarking using simulated data and gold-standard networks from model plants.
The project employs advanced machine learning techniques, network biology approaches, and computational statistics. The framework will be implemented in Python with GPU acceleration and tested on diverse plant expression datasets to ensure broad applicability across species.
About the PIs and their synergies
Dr. Peter Kindgren, an expert in experimental plant biology and lncRNA functional characterization, and Dr. Jian-Feng Mao, with extensive experience in plant genomics and bioinformatics method development, bring complementary expertise to this project. Dr. Kindgren's laboratory has successfully identified and functionally characterized key lncRNAs, while Dr. Mao has pioneered algorithms for integrating multi-omics data in complex biological networks. This research fills a critical gap between lncRNA identification and functional understanding, providing a powerful tool for systematic screening and hypothesis generation. The synergy between experimental validation expertise and computational innovation creates a bidirectional feedback loop, enabling comprehensive insights that neither investigator could achieve independently, making this project crucial for advancing plant molecular biology and agricultural applications.
