rSeqTU-A

rSeqTU-A predicts bacterial transcription units (TUs) from RNA-seq data using machine learning to delineate co-transcribed adjacent genes in prokaryotic genomes.

Key Features:

• Implementation: Implemented as an R package.
• Machine Learning Integration: Uses random forest for feature selection (including gene intergenic distance and continuous RNA-seq read count signals) and support vector machines (SVM) to build TU prediction models.
• Comprehensive Pipeline: Provides end-to-end processing including quality control of reads, read mapping, training set generation, feature selection, TU prediction, and visualization.
• Computational Compatibility: Supports execution on local computers and high-performance clusters.
• Enhanced Visualization Tools: Includes functionalities for visualizing predicted transcription units.

Scientific Applications:

• Transcriptional regulatory network analysis: Delineates transcriptional regulatory networks and dynamic regulatory mechanisms in prokaryotic genomes.
• Microbial genomics: Identifies co-transcribed gene clusters (TUs) to support genome annotation and comparative genomics.
• Transcriptomics: Facilitates interpretation of RNA-seq data to infer transcriptional organization.
• Systems biology: Provides TU-level information to inform systems biology models of bacterial gene regulation.

Methodology:

Performs quality control of reads, read mapping, training set generation, random forest-based feature selection using genomic (gene intergenic distance) and transcriptomic (continuous RNA-seq read count signals) features, and constructs support vector machine (SVM) models for TU prediction.