CStone

CStone assembles short-read RNA-Seq data de novo using de Bruijn graphs and classifies assembled contigs by underlying graph complexity to identify potential chimeric sequences.

Key Features:

• De Bruijn Graph-Based Assembly: Uses de Bruijn graphs and k-mer decomposition to assemble short-read RNA-Seq reads into contigs.
• Chimera Classification System: Assigns each contig one of three complexity levels based on the presence or absence of ambiguous paths to indicate potential chimerism.
• Complexity Annotation: Labels contigs with complexity levels to convey gene family complexity and ambiguous assembly regions.
• Benchmarking and Performance: Evaluated on simulated data (10 million read pairs from cDNA libraries of Drosophila melanogaster, Panthera pardus, Rattus norvegicus, and Serinus canaria) and on real datasets (~30 million read pairs from adult Drosophila melanogaster whole-body samples) with comparisons to Trinity and rnaSPAdes for contig length, sequence identity, and cDNA representation.

Scientific Applications:

• De novo transcriptome assembly: Generates assembled contigs from short-read RNA-Seq for species-agnostic transcriptomics analyses.
• Chimera detection and filtering: Identifies contigs likely to be chimeric via graph-complexity-based classification to improve assembly specificity.
• Gene family complexity assessment: Provides information on the range of gene family complexities present in datasets through contig complexity labels.
• Differential expression quality control: Informs on how chimeras within reference sets can impact identification of differentially expressed genes.
• Benchmarking and comparative evaluation: Facilitates comparison of assembly outcomes against Trinity and rnaSPAdes using both simulated and real RNA-Seq datasets.

Methodology:

Constructs de Bruijn graphs from short-read RNA-Seq, identifies contigs from the graph, and classifies contigs into three complexity levels based on graph topology and ambiguous paths.