ClusTrast

ClusTrast reconstructs transcript isoforms from short-read RNA-sequencing (RNA-seq) data using a de novo, cluster-guided assembly approach to produce comprehensive isoform sets for species lacking reliable reference genomes.

Key Features:

• Guided assembly: Employs a cluster-guided strategy that uses an initial primary assembly to generate guiding contigs for downstream steps.
• Primary assembly from short reads: Constructs an initial assembly directly from short-read RNA-seq data as the basis for clustering and guidance.
• Guiding contig clustering: Identifies and clusters a subset of guiding contigs to partition the assembly space for targeted processing.
• Clusterwise alignment and assembly: Aligns short reads to guiding contigs and assembles each clustered set of reads individually to improve reconstruction of complex transcript isoforms.
• Assembly merging: Integrates primary and clusterwise assemblies into a final, comprehensive transcriptome assembly.
• Comprehensive isoform recovery and low-expression recall: Produces a wide array of expressed known isoforms with high recall, including improved reconstruction at expression levels below the 15th percentile.
• Performance metrics: Reconstructs a significant proportion of reference transcripts to at least 95% of their length and identifies polymorphic variants in more than half of these transcripts, with reported superior performance relative to other de novo assemblers.

Scientific Applications:

• De novo transcriptome assembly in non-model organisms: Suited for assembling transcript isoforms in species without established reference genomes using RNA-seq short reads.
• Transcript-level gene expression analysis: Enables reconstruction of isoforms for downstream gene expression studies, including lowly expressed transcripts.
• Polymorphism detection at transcript level: Facilitates identification of polymorphic variants within reconstructed transcripts.
• Cross-species validation: Performance has been evaluated across datasets from six eukaryotic species.

Methodology:

Primary assembly construction from short-read RNA-seq data; clustering of guiding contigs from the primary assembly; alignment of short reads to guiding contigs and individual clusterwise assembly of each read cluster; merging of primary and clusterwise assemblies into a final transcriptome set.