MKMC enables reference-free transcriptomic analysis using k-mer representations

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Traditional RNA-sequencing (RNA-seq) relies on genome alignment and annotation, limiting its use in non-model organisms and potentially obscuring regulatory complexity.
• Alignment-based methods can introduce biases and may miss subtle regulatory variations.

Purpose of the Study:

• To introduce MKMC (Multi-sample Kmer Counter), a scalable, reference-free toolkit for RNA-seq analysis.
• To demonstrate MKMC's ability to detect biological variation and isoform-specific events without genome alignment.

Main Methods:

• MKMC utilizes k-mer-based statistics for analysis, bypassing the need for genome alignment.
• The toolkit integrates k-mer counting, abundance matrix generation, normalization, dimensionality reduction, and differential analysis.
• The approach was validated on diverse datasets, including killifish liver samples.

Main Results:

• MKMC successfully recapitulates known biological signals, such as sex differences in killifish.
• Its performance in differential expression analysis and transcriptomic age prediction matches alignment-based pipelines.
• MKMC identified and validated an unpredicted isoform-specific event, revealing hidden regulatory mechanisms.

Conclusions:

• MKMC provides a robust, extensible alternative to alignment-based RNA-seq analysis, enabling transcriptomic discovery in both model and non-model systems.
• The toolkit uncovers previously hidden isoform-level regulatory events contributing to transcriptional programs.
• MKMC is broadly applicable to k-mer-based analyses of next-generation sequencing data.