BinPacker

BinPacker assembles de novo transcriptomes from high-throughput RNA-seq reads by modeling isoform reconstruction as a series of bin-packing problems to resolve alternative splicing and coverage-driven isoform structures.

Key Features:

• Splicing junction-centric assembly: Assembly is performed by focusing on splicing junctions as the primary units for reconstructing isoform structures.
• Efficient handling of reads: A comb-like mechanism organizes RNA-seq reads along edges of a splicing graph to align reads with corresponding junctions for assembly.
• Integration of coverage information: Read coverage is incorporated into the model by representing item sizes and guiding placement within bin-packing formulations.
• Performance on real and simulated data: Demonstrated superior accuracy compared to nearly all existing de novo assemblers on both real and simulated RNA-seq datasets, including outperforming ab initio assemblers on a real dog dataset.
• Resource efficiency: Reported to run faster and use less memory than most competing de novo transcriptome assemblers.

Scientific Applications:

• Gene expression and regulation analysis: Reconstruction of transcript isoforms supports studies of gene expression and regulatory mechanisms.
• Novel transcript discovery: De novo assembly facilitates identification of previously unannotated transcripts from RNA-seq data.
• Alternative splicing analysis: Junction-centric reconstruction enables analysis of alternative splicing events and isoform diversity.
• Complex transcriptome assembly from short reads: Designed to assemble complex transcriptomes from short high-throughput RNA-seq reads.

Methodology:

BinPacker models assembly as tracking item trajectories where item sizes represent coverage and solves a series of bin-packing problems: bins represent potential isoforms, items are RNA-seq reads mapped to splicing junctions, placement is optimized based on coverage, and a comb-like mechanism arranges reads along splicing-graph edges.