LSTrAP

LSTrAP constructs co-expression networks from RNA-Seq data to enable genome-wide prediction and transfer of gene functions based on co-expression patterns.

Key Features:

• Integration of Essential Tools: Consolidates a suite of popular and high-performance methods into a cohesive workflow for constructing co-expression networks from RNA-Seq data.
• Parallel Computing Support: Optimized for parallel execution on computer cluster infrastructures and demonstrated on datasets of 876 Arabidopsis thaliana and 215 Sorghum bicolor samples.
• Quality Control Mechanisms: Includes robust quality-control procedures to identify and exclude spurious or low-quality samples.
• Functional Annotation Transfer: Constructs networks where nodes represent genes and edges connect significantly co-expressed genes to support prediction of gene functions by annotation transfer.
• Research Applications: Has been applied to group known photosynthesis-related genes in Sorghum bicolor and to suggest roles for previously uncharacterized genes.
• Implementation: Implemented in Python 3.4 or higher.

Scientific Applications:

• Gene function prediction: Uses genome-wide RNA-Seq co-expression to identify functionally related genes and assist annotation of uncharacterized genes.
• Functional genomics case study: Applied to group photosynthesis-related genes and propose functions for previously uncharacterized genes in Sorghum bicolor.

Methodology:

Constructs co-expression networks from RNA-Seq data, applies quality-control to exclude spurious or low-quality samples, and runs in parallel on compute clusters; networks represent genes as nodes and significant co-expression as edges.