RNA-Skim

RNA-Skim quantifies transcript-level abundances from RNA-Seq data using an alignment-free, sig-mer (k-mer) based approach.

Key Features:

• Alignment-Free Methodology: Eliminates read-to-reference alignment by using k-mer signatures for quantification.
• Partitioning into Transcript Clusters: Partitions the transcriptome into disjoint clusters based on sequence similarity.
• Introduction of Sig-Mers: Defines sig-mers as specific k-mers uniquely associated with each transcript cluster and uses them to inform abundance estimates.
• Parallel Processing Capability: Reduces the complex optimization problem into smaller tasks that can be processed independently and in parallel.
• Resource Efficiency: Operates using a small subset of k-mers (reported as <4%) and under 10% of the CPU time required by Sailfish, with >100-fold speedup versus alignment-based methods.
• High Accuracy: Produces transcript quantifications reported as comparable to or better than existing methods.

Scientific Applications:

• Gene expression studies: Provides transcript-level abundance estimates for analyses of gene expression.
• Differential expression analysis: Supplies quantifications suitable for comparing transcript abundances across conditions.
• Large-scale transcriptome profiling: Enables high-throughput transcript quantification for large datasets.

Methodology:

Partition the transcriptome into disjoint clusters by sequence similarity, identify sig-mers (specific k-mers uniquely associated with each cluster), use sig-mers to estimate transcript abundances within clusters while eliminating alignment, and reduce the optimization into smaller independent parallel tasks using a small subset of k-mers (reported <4%).